|
|
|
Sunday, 1 September 2019 | Dereel | Images for 1 September 2019 |
Top of page | ||
next day | ||
last day |
Blue Heelers: done!
|
Topic: multimedia, technology, opinion | Link here |
Finished downloading my episodes of Blue Heelers rather more quickly than I had expected. That was due in part to the fact that many of the episodes were in lower resolution:
01/S1-E05---Waiting-For-Apples-BLUH-005.mp4 VIDEO: [H264] 1280x720 24bpp 25.000 fps 2495.2 kbps (304.6 kbyte/s)
01/S1-E06---Apprehended-Violence-BLUH-006.mp4 VIDEO: [H264] 1280x720 24bpp 25.000 fps 2497.4 kbps (304.9 kbyte/s)
...
08/S8-E03---Dead-For-Quids-BLUH8-003.mp4 VIDEO: [H264] 896x504 24bpp 25.000 fps 1499.3 kbps (183.0 kbyte/s)
08/S8-E04---On-The-Road-BLUH8-004.mp4 VIDEO: [H264] 896x504 24bpp 25.000 fps 1497.0 kbps (182.7 kbyte/s)
So in the end I had “only” 349048 MB. And the numbers don't add up: according to the episode list, there should be 510 episodes, but I only have 508. In one case I know what's wrong. The download failed:
=== grog@teevee (/dev/pts/6) /spool/Series/Blue-Heelers/06 300 -> yt https://7plus.com.au/blue-heelers?episode-id=BLUH6-042
[7plus] BLUH6-042: Downloading JSON metadata
ERROR: No video formats found; please report this issue on https://yt-dl.org/bug . Make sure you are using the latest version; type youtube-dl -U to update. Be sure to call youtube-dl with the --verbose flag and include its complete output.
But why? Yes, I could report a bug, but why should only one download out of 509 fail? What about playing directly from the web browser? It plays the uninterruptible advertisements, but then hangs. What's it really trying to say? Downloaded the page and the previous one to compare:
=== grog@teevee (/dev/pts/6) /spool/Series/Blue-Heelers/06 297 -> fetch https://7plus.com.au/blue-heelers?episode-id=BLUH6-041
blue-heelers?episode-id=BLUH6-041 12 kB 111 MBps 00s=== grog@teevee (/dev/pts/6) /spool/Series/Blue-Heelers/06 298 -> fetch https://7plus.com.au/blue-heelers?episode-id=BLUH6-042
blue-heelers?episode-id=BLUH6-042 12 kB 111 MBps 00s=== grog@teevee (/dev/pts/6) /spool/Series/Blue-Heelers/06 299 -> diff -wu blue-heelers?episode-id=BLUH6-04*
--- blue-heelers?episode-id=BLUH6-041 2019-09-01 09:50:23.696833000 +1000
+++ blue-heelers?episode-id=BLUH6-042 2019-09-01 09:49:47.864950000 +1000
@@ -11,12 +11,12 @@
- <meta name="description" content="Ben becomes the centre of an emotional triangle when his ex-wife arrives unexpectedly, seeking a reconciliation, and a victim of domestic violence misconstrues his support for affection.">
- <meta property="og:description" content="Ben becomes the centre of an emotional triangle when his ex-wife arrives unexpectedly, seeking a reconciliation, and a victim of domestic violence misconstrues his support for affection.">
- <meta property="og:title" content="Blue Heelers - Married To The Job">
- <meta property="og:image" content="https://imageproxy.swm.digital/image?u=https%3A%2F%2Fimagemap.swm.digital%2Fimage%2FBLUH6-041&format=jpeg&w=1200&q=80">
+ <meta name="description" content="PJ is inconsolable when the Heelers discover Maggie's missing car in flames with a charred body inside.">
+ <meta property="og:description" content="PJ is inconsolable when the Heelers discover Maggie's missing car in flames with a charred body inside.">
+ <meta property="og:title" content="Blue Heelers - Web Of Lies">
+ <meta property="og:image" content="https://imageproxy.swm.digital/image?u=https%3A%2F%2Fimagemap.swm.digital%2Fimage%2FBLUH6-042&format=jpeg&w=1200&q=80">
<meta property="og:site" content="7plus">
- <meta property="og:url" content="https://7plus.com.au/blue-heelers?episode-id=BLUH6-041">
+ <meta property="og:url" content="https://7plus.com.au/blue-heelers?episode-id=BLUH6-042">
<link rel="preconnect" href="https://component-cdn.swm.digital/content" crossorigin>
@@ -72,7 +72,7 @@
<link rel="icon" href="/favicon.ico" type="image/x-icon" sizes="16x16">
- <link rel="canonical" href="https://7plus.com.au/blue-heelers?episode-id=BLUH6-041"/>
+ <link rel="canonical" href="https://7plus.com.au/blue-heelers?episode-id=BLUH6-042"/>
Clearly this isn't a youtube-dl problem. It must be the veil past which I could not see. But in fact, youtube-dl had a suggestion. At one point I mistyped a partial URL and got an interesting message:
=== grog@teevee (/dev/pts/6) /spool/Series/Blue-Heelers/06 300 -> yt blue-heelers?episode-id=BLUH6-042
ERROR: 'blue-heelers?episode-id=BLUH6-042' is not a valid URL. Set --default-search "ytsearch" (or run youtube-dl "ytsearch:blue-heelers?episode-id=BLUH6-042" ) to search YouTube
OK, that's worth trying:
=== grog@teevee (/dev/pts/6) /spool/Series/Blue-Heelers/06 301 -> youtube-dl "ytsearch:blue-heelers?episode-id=BLUH6-042"
[youtube:search] query "blue-heelers?episode-id=BLUH6-042": Downloading page 1
[download] Downloading playlist: blue-heelers?episode-id=BLUH6-042
[youtube:search] playlist blue-heelers?episode-id=BLUH6-042: Collected 1 video ids (downloading 1 of them)
[download] Downloading video 1 of 1
[youtube] _ejNUPiOUbA: Downloading webpage
[youtube] _ejNUPiOUbA: Downloading video info webpage
WARNING: Unable to extract video title
[youtube] _ejNUPiOUbA: Downloading MPD manifest
[dashsegments] Total fragments: 521
[download] Destination: _-_ejNUPiOUbA.f134.mp4
[download] 6.3% of ~103.08MiB at 1.03MiB/s ETA 03:48
How about that! What did I have?
-rw-r--r-- 2 grog home 894,381,099 29 Aug 09:17 S6-E40---Dirty-Money-BLUH6-040.mp4
-rw-r--r-- 2 grog home 894,360,821 29 Aug 09:38 S6-E41---Married-To-The-Job-BLUH6-041.mp4
-rw-r--r-- 1 grog home 150,764,279 1 Sep 09:56 _-_ejNUPiOUbA.mp4
That's rather less than the others. Took a look at it. The good news is that it does indeed appear to be a representation of the episode. The bad news is that the quality is terrible, and even worse, it only seems to be the top left quarter of the view. Still, it's intriguing. I wonder what it did there.
So that accounts for one of the missing episodes. What about the other? They've done everything they can to make it hard to find a missing episode: first, they have renumbered them (now 14 seasons instead of 13), and they have reallocated the episodes apparently at random. They also appear to have renumbered some of them, though that's not the issue here. So I numbered them sequentially to match the Wikipedia page:
=== grog@teevee (/dev/pts/6) /spool/Series/Blue-Heelers 335 -> k=1; for i in */*mp4; do echo -n $k: $i; echo; k=`expr $k + 1`; done | less
...
428: 11/S11-E40---Running-Scared-BLUH11-040.mp4
429: 12/S12-E01---Mind-Wide-Open-BLUH12-001.mp4
430: 12/S12-E02---On-The-Inside-BLUH12-002.mp4
431: 12/S12-E05---Secrets-and-Lies-BLUH12-003.mp4
According to the online list, there should be another episode between 428 and 429, “Reasonable Doubts”. Where is it? The detail page has the answer:
This episode of Blue Heelers went to air LIVE. ... NOTE: This episode is not included in the original DVD release. However, it is included in the ViaVision release as a special feature.
That explains it, but why doesn't Seven have it on their web site?
Understanding Trump supporters
|
Topic: politics, opinion | Link here |
Does Donald Trump lie? As we used to say years ago, “Is the Pope Polish?”. But recently I read a question on Quora: What do you think of China disputing that any calls took place regarding the trade wars with the United States over the weekend?. For me the answer is clear: Trump lied, maybe just to keep his hand in. But Stan Fasack answered differently (quoted verbatim, including punctuation):
Chinas stock market biggest loses in its history , growth slowest in two decades , needs it USA currency to purchase internationally , millions have lost their jobs , and companies are planning to move their manufacturing to a more friendly environment.
Then you ask the question which seems to dispute what trump has said ….
Guess it comes down to who do you believe ….. I choose our president , because it makes sense .
It takes all kinds, I suppose. I simply commented “… and because he would never tell a lie.”.
WRONG! I don't know what a lie is:
The clarion cry of the left ,truly falls on deaf ears of the 64 million American citizens who back trumps words and policies .
Trump exaggerates , so what ! But what is important is what trump delivers …. The left has so painted themselves into a corner that they think past rhetoric ,if repeated enough times would be believed by trumps base .
Here is the bad news for liberals, trumps base has grown , and three years of Russia Russia which had no legitimacy, has worn thin in trumps America .
Here is what IMO liberals never seem to understand, politicians have lied to mid Americans for decades , they lied to inner city minorities for decades , and educators have lied to all of America ,when we graduate young men and women who can not read or write at a high school level ,can’t do basic math , and have little chance of success in the real world.
Trump has made promises during his campaign to 64 million Americans , who believe two things about Trump … one is promises made promises kept , and the second is even the promises he has not delivered yet , he remains steadfast in trying to implement those policies and principals .
A final note , he has accomplished this without the help of a single democratic in congress ,
Later he made a further comment:
What you have misconstrued is what constitutes a lie ….. will give you two examples
When Obama said you can keep your doctor and you could keep your hospital , that was a lie ….why you may ask , well because he was president and he lied about a policy he initiated …
When trump says that unemployment is the lowest for minority inner city citizens , in all of recorded history…. That might not be accurate but it is true enough to be accepted ….. WH.Y. Because it is a positive outcome due to his policies …
Trump said for three years he did not collude with Russia to win election that was a true statement ,,, when Schiff and nadler said they had proof positive that he did , that was a lie .
Fascinating! It would be easy to put this down as just plain nonsense, but there's something very interesting in this answer and the response to my comment: an insight into how Trump can lie (my definition, thank you) as much as he wants and people will still not only support, but also defend him. It's also interesting that he considers it sufficient for Trump to support his supporters, and not all the US American people.
Power fail!
|
Topic: Stones Road house, general, opinion | Link here |
In mid-afternoon I heard a beep from the office UPS. That was nothing unusual when it was behind the flaky Eaton UPS in the garage, but now it's fed directly from the inverter. Something wrong there?
Yes! Power failure, continuing. OK, what do things look like? There was a mixture of overcast sky and shy sunshine, the battery had 63% charge, and the PV array was charging slightly. Nothing to worry about until things change. First, turn off the air conditioner: it could drain the battery in 90 minutes. And then do the cooking I had planned on gas rather than electricity.
Powercor was its usual confused self, reporting the fault multiple times with different causes. Durham Lead had two different faults, it seems, and I've decided that Buninyong-Mount Mercer Road is one of their favourite guesses at where the problem might be:
|
Can I hold out 3 hours? As it turned out, yes, no problem. The power actually came back at 17:20:03, by which time the battery state of charge had climbed to 70% and fallen back to 66%. But it was a good thing it happened when it did: it was getting dark, and I would soon have been dependent only on the battery. In fact, I had less than an hour's reserve:
The top line (bathtub) is the grid voltage, between 0 and 252 V, and the purple line is grid power consumption, up to 5 kW round 18:30. The dirty green area is the battery state of charge, which hit 20% at 18:16; without grid power it would have decreased to 5% by about 18:30, then shut down.
Summary: yes, the PV system saved me from an extended power outage. No, it didn't save me from nail-biting.
Monday, 2 September 2019 | Dereel → Ballarat → Dereel | Images for 2 September 2019 |
Top of page | ||
previous day | ||
next day | ||
last day |
First tulips of spring
|
Topic: gardening | Link here |
The first tulips are out:
|
That seems early, and indeed looking back over the last few years it's about 2 weeks earlier; normally they come pretty much exactly in mid-September:
|
|
|
The only exception was 3 years ago, when they were there in late August:
|
I suppose it must depend on the rainfall as much as anything, though it has (subjectively) been relatively mild.
Ashampoo fail
|
Topic: photography, technology, opinion | Link here |
Processing my photos involves putting it through Ashampoo photo optimizer. It often delivers good results, though I have had multiple issues with it over the years.
But then yesterday I had no results at all. After putting it through my comparison program, there was no obvious difference. OK, do it manually. Crash! “The photo could not be optimized correctly".
What caused that? Lots of experiments, and also bug reports, in which I established:
So what does it depend on? I had guessed some library, but that wouldn't explain the problems with euroa. At Tandem we once had a microcode bug that was triggered by an overflow in a (local) time counter, and we watched it start off in Australia, progress via India to Europe, and it would have gone on to the USA if we hadn't caught it first. Could it be something like that?
That's not the only bug with Ashampoo. In particular, it can't handle big images (somewhere round 50 MP seems to be its limit, and my E-M1 Mark II can generate 80 MP images). Time to look for something else, once again noting that “optimization” tends to mean “shrinking”, clearly not what Ashampoo 7 does. About the only thing I found was yet another Ashampoo version, the free “Optimizer 2019”. Downloaded that and ran into yet another Ashampoo bug that I have already seen but not recorded: entering the license key can take up to 30 minutes because of the slow echo on the input screen. Related to remote desktop, maybe? In any case, that was enough for today.
Kehr um!
|
Topic: general, music, opinion | Link here |
Off to pick up Yvonne this evening, the first time I've driven at night in quite a while, to the sounds of „Die schöne Müllerin“.
But something was wrong. The headlights seemed so dim. I was halfway down Grassy Gully Road as the singer (Peter Schreier?) sang „Kehr um!”, from part 15, „Eifersucht und Stolz“. Turn around! OK, stopped, turned around, went back home and changed cars. I can look at the cause of the dimness some other time, but I didn't want to discover that I had run out of electricity half way to Ballarat.
Who says that music has nothing to say?
Ballarat by night
|
Topic: photography, general, opinion | Link here |
Arrived in Ballarat deliberately a while before Yvonne's bus was due to take some night photos, something that I haven't done in a while. I was going to take the Leica Summilux 25 mm f/1.4 and the Leica DG Summilux 15 mm f/1.7 ASPH., but by coincidence I had received a newsletter from Olympus about... taking night photos. OK, worth reading, photos from different photographers. And they all recommended the M.Zuiko DIGITAL ED 7-14 mm f/2.8 PRO.
OK, I can do that too, so off with it instead. It was raining, and arguably some of the results didn't look that bad:
|
|
|
It's interesting also to note the fact that I was able to take photos hand-held with shutter speeds as slow as ½ s without any obvious camera shake. Tried some photos of car lights going past, like this one:
|
What puzzles me is that the car itself seems to have almost completely disappeared, not only in this photo.
Despite the rain, managed to keep the lens dry until the last 2 shots, showing the bus arriving:
|
Another thing that's really complicated is white balance. Which of these is correct?
|
|
|
I chose the first one, but the white lines on the street are clearly yellow. The second one, taking the white balance from the rego plate of the car on the left, is even yellower. But taking it from the road markings makes it unacceptably blue.
Außer Spesen nichts gewesen
|
Topic: animals, general, opinion | Link here |
Yvonne back more or less on time with a long story of her weekend. The course was Introduction to Trauma-Informed Horsemanship™ (admire that ™) with Sarah Schlote, who came from the other end of the Earth and had already caused me grief with payment.
Yvonne wasn't happy. She ended up with the feeling that Sarah really didn't know much about horses, and she was clearly also neither a good presenter—she kept repeating material—nor able to control the participants, who kept interrupting. And what she presented seemed to have little to do with the syllabus (PDF). In the end, apart from being $1000 or so poorer, Yvonne didn't have much to show for it. As the Germans say, „Außer Spesen nichts gewesen“, which doesn't translate well into English. Roughly: only expenses.
Tuesday, 3 September 2019 | Dereel | Images for 3 September 2019 |
Top of page | ||
previous day | ||
next day | ||
last day |
Echiums under way
|
Topic: gardening, opinion | Link here |
This time of year is fun: there's always something new coming into bloom. Today it was the Echium:
|
A replacement for Ashampoo?
|
Topic: photography, technology, opinion | Link here |
Got round finally installing Ashampoo Optimizer 2019 and getting past the license checks. It looks just like Optimizer 7: it seems that about the only difference from the older versions is the gratuitous change of the user interface. It works. But how about that, it, too, enlarges the images. But it has one difference from Optimizer 7: no batch mode! For that you have to pay money and upgrade to Optimizer 7. So that's a non-starter.
OK, I had some other package that looked like it might be able to do the same job. What was it... ah, right: movavi photo editor. Found out how to start it (why is it so difficult to find programs on Microsoft?), and sure enough, it didn't do too badly. But:
Before:
-rw-rw-r-- 3 grog lemis 2,916,842 3 Sep 14:49 Eurphorbia.jpeg
After:
-rwxr--r-- 1 grog wheel 14,540,829 3 Sep 17:02 Eurphorbia.jpeg
Once again the “optimized” version is nearly 5 times the size of the original. What causes that? Could it be that the images use different JPEG transforms from the ones that Ashampoo and movavi expect? I have some recollection of a base set of transforms (directory?) that each JPEG image uses, and potentially there's a conflict here. How can I find out? I really don't understand JPEG very well.
Air fried steak and kidney pie
|
Topic: food and drink, opinion | Link here |
Individual steak and kidney pies for dinner this evening. Normally we do this in the toaster oven (the real oven seems too energy-hungry), but today I tried it in the bigger of the two “air fryers”. I was just able to get the two forms into the device, but only because the bigger one is slightly oval:
How long? I started by pre-baking the bases, and then gave them 10 minutes at 200°. They didn't look bad at first:
|
But only the part in the middle of the fryer was properly browned; the outside was still raw. I had to turn it around and give it another 8 minutes at 180° (I think) to get a reasonable result:
|
I wonder if it's worth it. Like so many things with “air fryers”, it certainly doesn't scale.
Wednesday, 4 September 2019 | Dereel | Images for 4 September 2019 |
Top of page | ||
previous day | ||
next day | ||
last day |
More flash fun
|
Topic: photography, opinion | Link here |
Yesterday's flash photos of the steak and kidney pies worked well enough with the mecablitz 15 MS-1 ring flash, but on comparison discovered that the background was underexposed:
|
Why? Did a couple of comparisons, and once again the studio flash came out best:
|
But on reflection (well, consideration), that makes sense. The pots were correctly exposed, and since the background was further away, it got less exposure. That's a general limitation of on-camera flash, especially with larger distance ratios.
In the process of taking the photos, noted a further limitation of Olympus' remote flash: no information about the flash unit(s) in the Exif data. That's partially understandable, but it could at least report the flash unit on the camera.
Cleaning teevee
|
Topic: technology, opinion | Link here |
teevee has been making fan noises from time to time recently. Time to clean the dust out:
|
|
Put it back together, turned it on. Yes, the noise had changed. Now it was continuous. I had forgotten the graphics card, which took a lot of trouble to get out:
|
OK, more compressed air, ...
|
|
The second photo is a view of the heat sink under the fan. Not much difference. I had to go at it with cotton swabs before I finally got something worthwhile:
|
|
After that, no noise any more. Well, not until the evening, when it started up all over again. Time for a new graphics card, I think.
fsck!
|
Topic: technology, opinion | Link here |
Rebooting teevee wasn't the normal procedure I had expected:
|
Why? And didn't I have this on SUJ? Oh, no, I had had hangs with large updates with SUJ. But this should have been a clean shutdown. And there were a couple of interesting details in the recovery. Yes, under some circumstances you need to repeat fsck, and this was one of them. But then something caught my eye:
|
|
Why did the first pass say that the link count for the directory should have been 1? That's never a valid link count for a directory. It was empty, so 2 is correct (specifically, the directories . and ..). And once again I'm puzzled by the fact that such an old, dead directory should still be floating around.
More Ashampoo pain
|
Topic: photography, technology, opinion | Link here |
Many responses to my Ashampoo bug reports today:
I can't use version 7 because it pessimizes the file size. See my other tickets.
I'm sure that, once again, their bug tracking software is a big part of the problem. But surely it must be in their interest to provide better support than this.
Ashampoo 7 sizes: workaround
|
Topic: technology, photography, opinion | Link here |
More thinking. Why should Ashampoo Optimizer 7 create output files that are so much bigger than the original? Then it occurred to me: you can influence the size of JPEG images by setting the compression. How was it set on versions 7 and 2019? 100%! What nonsense, but at least it's a partial explanation. What happens if I set it to 75%? It's still larger, but almost acceptable.
So: I have reported this problem on multiple occasions. I'm told that only I have the problem, and that it must be my camera. It has (ostensibly) been reported to software development. Yet at least part of the problem is the ridiculous default compression that they set in their newer versions.
What can I say?
Daily NBN outage announcement
|
Topic: technology, general, opinion | Link here |
I wonder if there will ever come a day when no threat of an National Broadband Network routine outage exists. Today's came as usual:
NBNCo has let us know that they are planning network maintenance in your area, and that your service at <strong>29 STONES RD, DEREEL VIC</strong> will be affected.<br>
The details are:
- Start date and time: Fri 27th September 2019 00:00 AEST
- End date and time: Fri 27th September 2019 06:00 AEST
- Window: 6.0 hours
You may experience the following interruptions during the maintenance
- 1 min
The only surprise is the length of the outage. One minute, some time between 0:00 and 6:00. I can live with that, especially since they already have 8 hours of scheduled outages spread out over that whole day.
Drosera!
|
Topic: gardening | Link here |
Walking the dogs today, found these flowers:
|
Not many of them, but noticeable. It proves that they're some form of Drosera:
|
|
And I've never seen that many in flower at once.
Thursday, 5 September 2019 | Dereel | Images for 5 September 2019 |
Top of page | ||
previous day | ||
next day | ||
last day |
Another power failure
|
Topic: general | Link here |
Another short, harmless power failure this morning at 2:50:02.
Olympus high resolution again
|
Topic: photography, opinion | Link here |
While researching the Ashampoo problems, checked the high resolution mode on my Olympus OM-D E-M1 Mark II. I'm pretty sure that Ashampoo can't handle it.
Yes, that's the case. It seems that it can handle the JPEG output (just shy of 50 MP), but this native 80 MP image brings the message “Could not load the image. It is possible that the file format is not supported”:
|
How good is this high-resolution stuff? The first was taken with the E-M1 Mark II and the M.Zuiko Digital 45 mm f/1.8, at f/1.8. That should be fine from the sharpness point of view, but maybe not from the focus point of view, so I tried another with the M.Zuiko Digital ED 30 mm f/3.5 macro. Here a detail of a Hibiscus (of course), first at 20 MP, then at 80 MP:
|
|
Is that worth the trouble? Or should I try with something that is less likely to move between the individual images?
Ashampoo problems: solved!
|
Topic: technology, opinion | Link here |
Mail from a different person at Ashampoo today, with the usual plethora of attachments of previous messages, a total of 3175 lines. It started with:
Kay Hasbargen, Sep 4, 17:22 CEST
Dear Ashampoo customer,
We've recently received word about software issues from a few Ashampoo® Photo Optimizer 6 users.
We'd hate for you to experience similar issues, which is why we're happy to offer you a free upgrade to its successor!
Use Ashampoo® Photo Optimizer 7, worth $39.99, completely free of charge! This program supports Batch processing as well.
That was in reply to my message which included the line:
I would really appreciate it if you would read my responses.
And the next attachments, also from me (how do they order these things?):
On Wednesday, 4 September 2019 at 6:15:00 +0000, Patrick Bartsch (Ashampo=
o GmbH & Co. KG) wrote:
> Patrick Bartsch, Sep 4, 08:14 CEST
>
> do the recent problems also occur in Ashampoo Photo Optimizer 7?
I have already answered this question, and it's described in the link
I sent you. To repeat: no, but the output files are up to 7 times as
big as the original, making it useless.
As if that wasn't enough, as generous as the offer is, they should have noticed that I already have a license for Optimizer 7.
I'm left wondering how I would solve the organizational problems that Ashampoo has if I were in charge.
Ashampoo file sizes investigated
|
Topic: photography, technology, opinion | Link here |
Why are the output files from Ashampoo Photo Optimizer so big? The original 7 fold size increases have been explained: they were due to the 100% “quality”. But even when I select 75%, they're significantly bigger than the original. A poor choice of converter, perhaps? What happens if I create TIFF images, optimize them, and then convert them to JPEG with ImageMagick? Bingo! Much smaller images.
Is the quality really comparable? How can you tell? With lossy formats, you can't expect two images of the same quality to be identical.
More spring flowers
|
Topic: gardening | Link here |
Walked through the forest with the dogs today. The Drosera are flowering there too, though not as “profusely”:
|
|
|
Also Pterostylis:
|
|
An unidentified wildflower, about 1 cm across:
|
And lots of different Acacias:
|
|
|
|
|
Garden work and dead plants
|
Topic: gardening, opinion | Link here |
Finally got round to pruning our « Monsieur Tillier » rose, removing probably 80% of the growth. Also a chat with Garry and Diane Marriott from next door. They're planting lots of plants along the border between the properties, apparently not for the first time. They, too, have experienced the remarkable plant death that I've been reporting over the last 4 years. As if to make the point, Diane gave me another lilac plant to replace the many that have died. I wonder what's wrong with the soil round here.
Friday, 6 September 2019 | Dereel | Images for 6 September 2019 |
Top of page | ||
previous day | ||
next day | ||
last day |
Ashampoo bug search
|
Topic: photography, technology, opinion | Link here |
Another email from Ashampoo this morning:
Date: Thu, 05 Sep 2019 11:47:17 +0000
please send an example picture (original before optimization and optimized after optimization) so we can take care of it.
What a good idea! I should have had it myself. In fact, I did:
Date: Mon, 02 Sep 2019 07:15:47 +0000
I don't understand why this has happened. I haven't changed any software or settings. I'm attaching one image in JPEG form (output from DxO PhotoLab). The original raw image is available, but your software doesn't like the size.
But you can't really blame them for that. Their bug tracking system is so terrible that nobody can find anything in it. But if my assumptions are correct, this should happen with any image. And it does. Tried it with the sample images they supply, and got:
-rwx------ 1 grog lemis 1,085,839 16 May 15:07 optimized/statue-batu-caves.jpg
-rwxr--r-- 1 grog lemis 773,256 16 May 15:07 originals/statue-batu-caves.jpg
Now admittedly that's not as bad as my experience, but it should give them something to go on. But why do they have to ask me about this?
In passing, how horrible is Microsoft file system navigation. It took me 15 minutes to find those files. The original proved to be in C:\Users\Public\Pictures\Ashampoo and the result ended up on eureka's /Photos/Ashampoo_grog/photoOptimizeHistoryDataBase. How did it get there? This was a virgin installation of Ashampoo Optimizer 7, but clearly it sucked in knowledge from version 2016 without so much as a by-your-leave.
Summer meets spring
|
Topic: gardening | Link here |
Flowers in the lounge room today:
|
|
The roses (admittedly, not the prettiest, nor the biggest) are from last summer, and the daffodils are, of course, fresh.
Saturday, 7 September 2019 | Dereel | Images for 7 September 2019 |
Top of page | ||
previous day | ||
next day | ||
last day |
Improving JPEG size
|
Topic: photography, technology, opinion | Link here |
After my experience of the last few days, spent some time today converting my photo processing workflow from using JPEG intermediate images to using TIFF. A couple of surprises:
The TIFF output from DxO PhotoLab contains two images: a full-size image and a really tiny thumbnail:
=== grog@eureka (/dev/pts/13) ~/Photos/20190906 390 -> identify Daffodils-Roses.tiff
Daffodils-Roses.tiff[0] TIFF 5184x3888 5184x3888+0+0 16-bit sRGB 121.2MB 0.008u 0:00.125
Daffodils-Roses.tiff[1] TIFF 162x121 162x121+0+0 8-bit sRGB 121.2MB 0.000u 0:00.000
To convert the image to JPEG, I need to specify which image. From the Makefile:
# Make JPEGs for any TIFFs in the main directory
JPEG:
for i in *.tiff; do \
MYJPEG=`basename $$i tiff`jpeg; \
if [ ! -e "$$MYJPEG" ]; then \
convert -quality 85% $$i[0] $$MYJPEG; \
${PHOTOTOOLS}/exifcopy $$i $$MYJPEG; \
fi; \
done
It's the [0] that selects the image.
Of course, it's not said that all TIFFs that I encounter will have images in this order. Probably the correct thing to do is to investigate the contents and choose the biggest one.
My contacts target (creating a web page with “contact prints”) has problems: it links to the original image. When that was a JPEG, web browsers could display it. But they can't do anything with TIFFs, and the effort to convert Every Image to TIFF is probably not worth it. But if it is, I know how to do it.
Investigating Ashampoo compression
|
Topic: photography, technology, opinion | Link here |
So how do the file sizes of Ashampoo Optimizer compare with Optimizer 6 when using the same JPEG compression setting? I can't convert with Optimizer 6, but I have plenty of images which have been converted, at the rather unusual default setting of 92%. Why not take the one I complained about six months ago? There I had:
=== grog@eureka (/dev/pts/11) ~/Photos/20190316 89 -> l _Ashampoo_Photo_Optimizer_Backup/Hellebore-2.jpeg Hellebore-2.jpeg
OK, take that image (the unoptimized one) and put it through Optimizer 7 at different compression rates. Here the results.
Original from DxO (85%): -rwxr--r-- 1 grog lemis 4,323,518 16 Mar 17:29 Hellebore-2.jpeg
Optimizer 6, 92% -rwxrw-r-- 2 grog lemis 6,632,721 16 Mar 17:29 Hellebore-2.jpeg
Optimizer 7, 100% -rwxrw-r-- 2 grog lemis 23,475,024 16 Mar 17:29 Hellebore-2.jpeg
Optimizer 7, 92% -rwxr--r-- 1 grog lemis 6,973,661 7 Sep 16:23 Hellebore-2.jpeg
Optimizer 7, 85% -rwxr--r-- 1 grog lemis 5,737,038 7 Sep 16:37 Hellebore-2.jpeg
Optimizer 7, 75% -rwxr--r-- 1 grog lemis 4,700,721 7 Sep 16:21 Hellebore-2.jpeg
ImageMagick 100% -rw-r--r-- 1 grog lemis 28,018,272 7 Sep 16:28 43166663_DxO.jpeg
ImageMagick 92% -rw-r--r-- 1 grog lemis 8,072,150 7 Sep 16:29 43166663_DxO.jpeg
ImageMagick 85% -rw-r--r-- 1 grog lemis 4,207,519 7 Sep 16:29 43166663_DxO.jpeg
ImageMagick 75% -rw-r--r-- 1 grog lemis 2,995,312 7 Sep 16:30 43166663_DxO.jpeg
The Ashampoo images were optimization of the original 4.3 MB image. I can only compare Optimizer 6 at 92%; there Optimizer 7 is roughly 5% larger than Optimizer 6, and 61% larger than the original. Even at 75% the result is larger than the original.
The ImageMagick results aren't directly comparable, nor should they be. They reflect the alternative that I have chosen: convert to TIFF with DxO, optimize it with Optimizer 7 (producing a TIFF output file) and then convert that TIFF to JPEG with ImageMagick. The results are interesting: at 100%, the image is significantly larger than the Optimizer 7 output from JPEG; but the sizes diminish rapidly as the quality decreases. I still need to think this one through.
The other issue is that Ashampoo claimed that the problem appeared to be due to my camera. OK, what's a popular DSLR? Canon 80D? Let's look at some sample photos from that camera. Digital Photography Review has galleries, in this case here. Took a look and found this one:
It's interesting because it clearly needs optimization; in fact, the DPReview people did one themselves with Adobe Camera Raw:
OK, download JPEG and raw image, convert the latter with DxO PhotoLab, optimize:
-rw-r--r-- 2 grog lemis 10,452,866 24 Mar 2016 0370580967.jpg
-rwxr--r-- 1 grog lemis 18,406,050 7 Sep 16:53 5922129759-100%.jpg
-rwxr--r-- 1 grog lemis 2,922,363 7 Sep 16:58 5922129759-75%.jpg
-rwxr--r-- 1 grog lemis 4,075,477 7 Sep 16:45 5922129759-85%.jpg
-rwxr--r-- 1 grog lemis 8,674,434 7 Sep 16:56 5922129759.jpg
-rwxr--r-- 1 grog lemis 25,016,811 7 Sep 16:54 Dan-Canon80D-22_DxO-100%.jpg
-rwxr--r-- 1 grog lemis 4,742,910 7 Sep 16:57 Dan-Canon80D-22_DxO-75%.jpg
-rwxr--r-- 1 grog lemis 6,079,607 7 Sep 16:45 Dan-Canon80D-22_DxO-85%.jpg
-rwxr--r-- 1 grog lemis 5,875,583 7 Sep 16:55 Dan-Canon80D-22_DxO.jpg
5922129759.jpg was the original downloaded JPEG, and Dan-Canon80D-22_DxO.jpg the converted raw image (at 85%). The difference in size is surprising, as is the difference between the original JPEG and the one optimized with Adobe Camera Raw (10.5 MB). But one thing is clear: here, too, the images are much larger. Clearly Ashampoo can't have tested their software very well.
In passing, once again I note how hard it is to live with Microsoft. How do I delete the old images to start again? It seems that I have to eliminate the entire hierarchy, otherwise Ashampoo loads all potential images, including the backups, which have the same file name and appearance, and there's no way to tell them apart.
And apart from investigations, how did Ashampoo's automatic Optimizer compare with DPReview's manual use of Adobe Camera Raw? Here are the comparisons, first the out-of-camera JPEG, then the output of DxO PhotoLab, then the Ashampoo optimization, then the manual optimization (run the cursor over an image to compare it with its neighbour):
|
|
|
|
Clearly it's one of the last two (Ashampoo or ACR). I asked Yvonne, and she preferred the Ashampoo version. I'm wondering if it isn't a little too green.
About the images from DPReview: I've contacted them and asked for confirmation that it's fair use. Potentially, though, they will ask me to remove them, in which case only the text will remain.
Cailles aux raisins again
|
Topic: food and drink, opinion | Link here |
It's been pretty much exactly 2 years since we last ate caillles aux raisins. Yvonne no longer likes the idea, but surely there must be a way to get rid of the 10 quails that we still have in the deep freeze.
Minor tweaks this time: more brandy, bacon over the quails, add a shallot (these last two both from the cailles aux raisins secs), and more grapes. The results were OK. I didn't change it today, but I've decided that the 30 minute cooking time isn't enough. I'll try 45 minutes next time.
Blog admirer
|
Topic: general, opinion | Link here |
We had just sat down for dinner when my mobile phone rang in the office. That's unusual. So far no spammers have found it. It's also unusual that I heard it: on my home page I write:
Since getting the second radiation tower in Dereel, we now have coverage, and the phone is usually turned on. But I don't carry it with me, and I won't hear if it rings. Use only if I tell you that it will be answered.
Who was it? I don't know. I have the phone number, of course, but whoever it was didn't give his name—possibly I didn't answer in an overly friendly tone, but it wasn't that unfriendly either. “Just wanted to tell you that I love your blog”. “Thanks”. And then he hung up.
Who was it? Nobody who read my diary, you'd think, if he refers to blogs. But the blog is 13 years out of date, so maybe he hasn't read the bottom, where I say:
This is a diary, not a “blog”
I don't have that many admirers, and I'm left wondering whether I didn't come across as overly unfriendly. The time certainly didn't help.
Sunday, 8 September 2019 | Dereel | Images for 8 September 2019 |
Top of page | ||
previous day | ||
next day | ||
last day |
Flash problems after all
|
Topic: photography, food and drink, opinion | Link here |
Mee soup for breakfast. Yvonne thought it looked good, and that I should take a photo.
She soon wished that she hadn't. Things didn't go quite as planned. I grabbed my trusty mecablitz 15 MS-1 ring flash, put it on the M.Zuiko Digital ED 12-100 mm f/4.0 IS PRO , and took a photo:
|
What's that black ring? And that shadow? The ring was clear: vignetting. The flash extends so far in front of the lens that it obstructs it at focal lengths less than about 16 mm:
|
OK, longer focal lengths, here 25 mm:
|
That shadow is still there! Is that available light from the window? No, it's clearly coming from above, while the window was on the left. To be sure, after breakfast I tried once with available light, and once with flash and 1/250 s exposure:
|
|
Not only is the shadow different (and less pronounced this time), the exposure is 5.6 EV less.
Is it the position of the flash tubes? Tried rotating the thing in increments of 90°, to no avail, and in any case, I later confirmed that the flash tubes are where you'd expect:
|
In passing, it's interesting that you can take this kind of photo and get correct exposure. At least that works. And in this case there was no issue with the shadow.
Different lens? I tried with the M.Zuiko Digital ED 30 mm f/3.5 macro, again with no clear explanation. These two should have been identical:
|
|
It's clear that the shadow must have something to do with the flash unit itself. But it's always at the bottom. And how could it take up half the image, like in the image immediately above? It can't be flash sync problems; they're very different in shape, perfectly straight, and they also come down from the top of the image, not up from the bottom. It can't be the lens, since it happened with two different lenses, though not the Leica DG Vario-Elmarit 12-60 mm f/2.8-4. I suspect that that's a coincidence.
I'm baffled.
Another power failure
|
Topic: general, opinion | Link here |
Short 2 second power failure this afternoon at 13:07:03.
Monday, 9 September 2019 | Dereel | Images for 9 September 2019 |
Top of page | ||
previous day | ||
next day | ||
last day |
Ring flash vignetting compared
|
Topic: photography, opinion | Link here |
I'm still puzzling about yesterday's flash vignetting issues with the mecablitz 15 MS-1, and at some point I'm going to have to do some more experiments, once I have thought things through.
The big puzzle is the shadows at the bottom of the image:
|
I still have no idea what that could be. Camera strap? No, apart from the fact that it's too repeatable, I didn't have a strap on the camera. It was still mounted on the panorama rail from Saturday. The rail itself? No, it was behind the flash, and I had to remove it to mount the shorter macro lens, but the shadow remained.
Other comments on vignetting on the web? No, couldn't find anything useful. I wonder how many of these flashes Metz sold.
The other issue, the real vignetting, is more obvious. How do the flashes compare? I have four ring-like macro lights: the “Viltrox JY-670 Macro Ring Lite” (now disowned by its makers web site), the Olympus STF-8, the toy LED flash and the mecablitz. How do they compare? Here on the E-M1 with the Leica DG Vario-Elmarit 12-60 mm f/2.8-4, which exhibits similar vignetting to the M.Zuiko Digital ED 12-100 mm f/4.0 IS PRO:
|
|
|
|
From this point of view (and really only from this point of view), the STF-8 is the clear winner. But it took me 5 minutes to attach. It's really a horrible unit. And clearly the mecablitz poses the biggest problems. Why didn't they put the mount further forward? Might that make it more difficult to attach?
The eternal Brexit
|
Topic: politics, language, opinion | Link here |
Brexit continues to fascinate from numerous viewpoints. The good news is that British democracy is more resilient than the US American democracy. But the activities of the last week are still mind-boggling:
Boris Johnson loses the parliamentary majority for his party, narrowly not because of the defection of his own brother Jo, who waited to take second place.
21 of his own party members voted against him in a critical vote. What do you do if you're in a hole? If you're Boris Johnson, it seems, you kick them out of the party (or, to use a term that most of the world has heard for the first time, withdrew their whip, which, it seems, means party membership). With another couple of protest resignations, that's round 25 below majority; I've lost count.
The vote was to avoid a “no-deal Brexit”, which was looking more and more likely, especially as Johnson appears to have done nothing to improve the chances of a deal. It went through the House of Lords on Friday and was approved. In the meantime, Johnson stated that he would “rather be dead in a ditch” than ask the EU for an extension.
And that would put him in contempt of parliament, with the threat of imprisonment if he didn't comply. On the other hand, he has also ruled out resignation.
Today the Queen has approved the bill for the extension. And, as promised, parliament has been prorogued (there's another word nobody knew) until 14 October, at a session where Johnson was not present. They're certainly not going down without a fight:
Johnson's future hinges on him getting a “deal” Brexit by 19 October. Given the turmoil of last week (six lost votes in six days), I can see lots of fun and games happening between 14 and 19 October.
People compare Johnson with Trump, some even disfavourably. But what they have overlooked is that the British parliamentary system, maybe because it's not cast in concrete like the US Constitution, looks like it will weather the tide better.
How will it end? Here's a rather unlikely possibility from Reddit:
Tuesday, 10 September 2019 | Dereel | Images for 10 September 2019 |
Top of page | ||
previous day | ||
next day | ||
last day |
Power fail!
|
Topic: Stones Road house, general, opinion | Link here |
Up this morning to discover the power out to the microwave ovens. But the main oven was still on. RCD trip! And this time not the UPS. OK, turn on again. But that meant that eureka was down too!
Damn! What's wrong with this house? At the very least it means that I really need to get the wiring updated. Out with the two old RCDs, put one on every individual circuit instead.
When did the failure happen? My power statistics database gives me one indirect indication: some time before 01:35:24, when the office UPS gave out.
Identifying USB disk drives
|
Topic: technology, opinion | Link here |
I have a number of external backup disks on eureka: one for eureka itself (/backups), one for the other computers on the network (/dump), one for photos (/photobackup) and one for videos (/videobackup):
|
When eureka came up again, it identified the external backup disks in the wrong sequence. They come up as SCSI disks (/dev/da0, for example), but the numbering seems to change on every boot. The only way I can find to fix things (apart from editing /etc/fstab) is to disconnect all the ones including and after the first one that has been misnumbered. Today I finally got round to labeling them all, in the process discovering one that I had forgotten (the one just visible under the power point on the right):
|
What's that? It proved not to be plugged in to the USB hub, but when I did, all I got was:
Sep 11 09:36:57 eureka kernel: usb_alloc_device: set address 15 failed (USB_ERR_TIMEOUT, ignored)
Sep 11 09:36:57 eureka kernel: usbd_setup_device_desc: getting device descriptor at addr 15 failed, USB_ERR_TIMEOUT
Sep 11 09:36:58 eureka kernel: usbd_req_re_enumerate: addr=15, set address failed! (USB_ERR_TIMEOUT, ignored)
...
Sep 11 09:37:05 eureka kernel: ugen0.15: <Unknown> at usbus0 (disconnected)
Sep 11 09:37:05 eureka kernel: uhub_reattach_port: could not allocate new device
What's broken there? Cable? Disk? Hub? I'll check some other time.
Sublimed ice
|
Topic: food and drink, opinion | Link here |
Tomato soup this evening, from the freezer. It had been in there for a couple of years, and it showed:
|
How did that happen? It seems that the ice from the soup sublimed and crystallized above the block of soup. Under what circumstances can that happen?
More vignetting issues
|
Topic: photography, opinion | Link here |
Took the photos of the sublime soup with the mecablitz 15 MS-1 ring flash again, this time with the M.Zuiko Digital ED 14-150 mm f/4.0-5.6 lens, which has a smaller front element. Vignetting? Yes:
|
That was at a focal length of 14 mm, corresponding quite closely to the limits with other lenses. It was gone by about 18 mm, but again I had the shadow at the bottom.
|
I'm completely baffled about what causes it. It's not the shape or the orientation of the ring flash: in one of these images I rotated the flash by 90°, but the shadow remained the same. But it seems that, like the real vignetting, it goes away at longer focal lengths, round 50 mm (100 mm “full frame” equivalent). That's presumably why I didn't see it with the M.Zuiko Digital ED 60 mm f/2.8 Macro. But the more I examine this issue, the more puzzled I become.
Configuring w3 web site
|
Topic: technology | Link here |
For some reason I have been dragging my heels on configuring the new web site. I have until the end of the month for one deadline, but the fact is that the current site is completely overloaded, and I really need to do something.
Did the first step today: straight Apache configuration. That was easier than I thought. I had some recollection of putting all web site descriptions in subdirectories, but the instructions I had (maybe almost as obsolete as The Complete FreeBSD) just modified the main http.conf file:
--- httpd.conf 2019/08/22 04:02:47 1.1
+++ httpd.conf 2019/09/10 06:05:46
@@ -178,6 +178,7 @@
#LoadModule userdir_module libexec/apache24/mod_userdir.so
LoadModule alias_module libexec/apache24/mod_alias.so
#LoadModule rewrite_module libexec/apache24/mod_rewrite.so
+LoadModule php7_module libexec/apache24/libphp7.so
# Third party modules
IncludeOptional etc/apache24/modules.d/[0-9][0-9][0-9]_*.conf
@@ -213,7 +214,7 @@
# e-mailed. This address appears on some server-generated pages, such
# as error documents. e.g. admin@your-domain.com
#
-ServerAdmin you@example.com
+ServerAdmin grog@lemis.com
#
# ServerName gives the name and port that the server uses to identify itself.
@@ -222,7 +223,7 @@
#
# If your host doesn't have a registered DNS name, enter its IP address here.
#
-#ServerName www.example.com:80
+ServerName www.lemis.com:80
#
# Deny access to the entirety of your server's filesystem. You must
@@ -246,8 +247,10 @@
# documents. By default, all requests are taken from this directory, but
# symbolic links and aliases may be used to point to other locations.
#
-DocumentRoot "/usr/local/www/apache24/data"
-<Directory "/usr/local/www/apache24/data">
+# DocumentRoot "/usr/local/www/data"
+# <Directory "/usr/local/www/data">
+DocumentRoot "/home/grog/www.lemis.com"
+<Directory "/home/grog/www.lemis.com">
#
# Possible values for the Options directive are "None", "All",
# or any combination of:
No surprises there, but also no PHP support. That'll be the next step, and there I'm expecting more problems, including PHP's habit of forgetting its history, requiring me to rewrite at least some of my 10,000 lines of PHP code.
Smart devices: the pain
|
Topic: technology, language, opinion | Link here |
HowToGeek is full of articles about Smart Homes, including information about how to work around the complete lack of standardization in the area, but without explaining that it's a problem, or explaining why you would rather say “Light in lounge room: off” rather than just turning the thing off. In general I'm still amazed by the amount of attention given to voice commands. Maybe it's an indication that people can no longer type, either because they never learnt, or because they no longer have something to type on.
Yvonne came up with this illustration of the problems of voice input:
Wednesday, 11 September 2019 | Dereel | Images for 11 September 2019 |
Top of page | ||
previous day | ||
next day | ||
last day |
Another RCD trip!
|
Topic: general, opinion | Link here |
Yvonne out early this morning, in principle leaving me to sleep. But she couldn't: the RCD for most house power had tripped again, taking eureka with it. Why? This time power went out before 5:20:09. Obviously I needed to change back to the UPS circuits, which haven't failed since I reset it. But high time to get things fixed.
More flash pain
|
Topic: photography, opinion | Link here |
Taking the photos of the flash units behind the monitors was an issue: which flash unit should I use? I didn't even bother to try the mecablitz 15 MS-1. OK, put up with the pain and use the Olympus STF-8:
|
That's not very impressive. What went wrong? That was at f/7.1, at which aperture I should have had a range of round 90 cm. Even this flash unit should have managed that. Still more playing around. Adjust flash exposure compensation? +1 EV? +2 EV? No difference! Open aperture? f/4? f/3.6? Yes, gradually it's better:
|
Finally turned up the sensitivity to 800/40° ISO, where things got better, and turned the compensation back to 0 again (it really seems to make no difference whatsoever):
|
At least the exposure was better, but somehow the image gradation doesn't seem to be. Why is this all so difficult?
In summary: the STF-8 is too weak for even close quarters, and under some circumstances exposure compensation doesn't work. Is that because I had remote flash selected? It certainly provides for the compensation, so it should work.
Riders pass
|
Topic: general, animals | Link here |
Round lunch time Yvonne saw a large group of horse riders pass the house. They certainly left their mark on the freshly graded road:
|
|
And one of their horses had the most enormous hooves that we have seen:
|
Spent some time discussing whether the horse had a hoof problem, based on the unevenness and angle of the hoofprints.
Beef skewers, air fried
|
Topic: food and drink | Link here |
Yvonne brought some beef skewers kebabs back
today. More fun in the bigger air fryer.
It wasn't until later that I discovered that I had done this before. I gave them 14 minutes at 230°. OK, but really a little overdone.
Next time I'll do 10 minutes, like last time, but at 230°.
Thursday, 12 September 2019 | Dereel | Images for 12 September 2019 |
Top of page | ||
previous day | ||
next day | ||
last day |
Power fail!
|
Topic: general | Link here |
My RCDs and UPS aren't the only cause of power failures, of course. After the fun we have had locally in recent days, it seems that Powercor had to get on the scene, with no less than 6 failures in an hour. The longest was 20 seconds, and of course I didn't find out until I got my daily report. That's been 28 failures in 5 months, well above the average, and 0 failures that affected house power, infinitely below the average.
Flash shadow problems: solved!
|
Topic: photography, opinion | Link here |
Mail from Tim Bishop today, pointing out something that I hadn't seen on one of my flash test photos:
|
I took this photo (and a few other photos) specifically to look at where the flash was coming from, and saw that the ring flash was uniform on both sides. But I completely missed that the on-camera flash also fired! It wasn't supposed to do that: it was set to “OFF”. But that would completely explain the shadow, and also that it didn't help to rotate the flash unit.
Back to the bathroom to take some more photos:
|
|
The background is dark because I set the camera to flash only, and that's what I expected. But why the difference in the on-camera flash? In the second photo I had it pointing upwards. And sure enough, like that the shadows were gone, even with the extreme vignetting at 12 mm:
|
|
The shadow above the dispenser in the second image is because the flash tubes are mounted on the sides, and thus to be expected. OK, but why is the on-camera flash firing? Did I have the settings wrong? Maybe I did, but they're the way I expected them to work.
Is it maybe a limitation of the toy flash that I was using? Tried with the mecablitz 58 AF-2 O digital. And while I was at it, why use TTL flash? Using manual at lowest power level (1/256) should do the trick.
What a pain it is to set these things! I had to count down a total of 24 power levels, one button press at a time, to get there. I couldn't even wrap around. And when I was done, I had:
|
No flash at all from the ring flash! Somewhere there are communication issues that the documentation is not explaining. Is this maybe some subtle change since the introduction of mirrorless cameras? The Olympus DSLRs had an infrared emitter that might be part of the secret. Tried it on the E-30, but it refused to work at all unless the pop-up flash was up. Interestingly, it didn't have the fourth line for what I thought was the on-camera flash (second image):
|
|
So it didn't seem worth following up on that one. I'll have to research more (or wait for another Tim Bishop to come along and tell me what I'm doing wrong).
Armed with that success, tried yesterday's problem with the disk photos. Here the first attempt with the Olympus STF-8 (first) and the 15 MS-1:
|
|
Strange. Both were underexposed. Is this an issue with the camera itself, something like highlight priority (the power point on the right)? I've had issues in the past that suggest that the camera isn't very clever in measuring TTL flash exposure.
OK, what does the 15 MS-1 do with exposure compensation? Here the same two with +2 EV compensation:
|
|
Curiouser and curiouser! Exposure compensation works as expected with the mecablitz, but with Olympus' own flash it doesn't!
Did a bit of investigation, including finding a couple of YouTube videos that didn't really tell me very much that I didn't know. They did point to firmware updates, including a PDF document, describing the firmware changes, which proved to be half a page in size with no particular markup. The latest release is 5.0, and that's what I have, so there's no issue there.
So where am I?
Friday, 13 September 2019 | Dereel | Images for 13 September 2019 |
Top of page | ||
previous day | ||
next day | ||
last day |
Photos of switchboard
|
Topic: photography, general, opinion | Link here |
It's clear that it's time to document the power distribution in the house. The switchboard is full of warnings, but there's no clarity how things are interconnected:
|
How do I do it right? But more to the point, how do I get a good photo of the switchboard? The dark line above is a cover, and it doesn't stay in that position when I lift it:
|
My first attempt was with a rod on either end. Here the first of two:
|
Then I stitched them with Hugin, with less than perfect results:
|
That's not out of focus, just two images not quite overlapping. And apart from the bad registration, I had taken it with direct flash, and the labels had reflected. I could probably fix the overlapping with manual choice of control points, but not the reflections. Repeat the attempt with bounce flash and Yvonne standing in for the rod. But the angle of the lid was critical, and I wasn't able to get the images to align properly:
|
Geriatric US president?
|
Topic: politics, opinion | Link here |
When Donald Trump became US president, one of the many criticisms was that he was so old, over 70 years at his inauguration. That makes him the oldest president in the history of the USA at his inauguration. Give him two terms (no, please, don't!) and he would be 78 by the end of it.
But hopefully he won't get reelected. Who would be the new president then? The current forerunners are Bernie Sanders, Joe Biden and Elizabeth Warren. And they're all older than Trump! On the inauguration day, 20 January 2021, Bernie Sanders will be 79, Joe Biden will be 78, and Elizabeth Warren, the youngster in the group, will be 71. Imagine a two-term Sanders leaving the presidency at the ripe old age of 87! The oldest so far was Ronald Reagan with not quite 78.
Can't they come up with younger people to run their country?
More mystery flowers
|
Topic: gardening, opinion | Link here |
While walking the dogs this afternoon, saw a couple of flowers of interest:
|
|
The first looked like a Kniphofia, but it's far too small, and it's flowering too early in the year. It also doesn't have the typical cone of a Kniphofia.
And the second one suggested Drosera, but it's growing in water, and I can't see the rest of the plant anywhere.
Margaret Swan again
|
Topic: food and drink, general | Link here |
Margaret Swan is in town, and she and Chris Bahlo came along for dinner. It's always good when Margaret comes, because I can cook Indian food. Spent most of the afternoon doing that, and for some reason people particularly liked it this time.
Saturday, 14 September 2019 | Dereel | Images for 14 September 2019 |
Top of page | ||
previous day | ||
next day | ||
last day |
Democracy and populism
|
Topic: politics, history, language, opinion | Link here |
Western societies have been democratic for centuries. Or have they? Democracy is government by the people. They tried it in ancient Athens, only possible because the city was swimming in money, and decided that it was a bad idea.
What we have now, at least in most places, is a representative democracy, an indirect democracy: people elect somebody to represent them in parliament (or whatever the country chooses to call it). But who elects? Even today, children under 18 are not allowed to vote in most countries. And the very first place where women were allowed to vote (perish the thought!) was New Zealand, only 126 years ago. The Swiss canton of Appenzell Innerrhoden didn't introduce personal voting until 1991.
People highlight voting rights for women (and, in the USA, “people of color”), but it's really only been in the last 200 years that most men have been allowed to vote. And many representatives seem to have lost track of the fact that they're just that, representatives (of the people in their constituency). Enter the populists, promising the electorate what they want (but possibly not what they need). “Populism” has become a dirty word.
But really, it's just Latin for “democracy”, or at least something very similar. Greek δῆμος means the common people, just like populus in Latin. So really the populists are the real democrats.
What does that say about the populists? I don't know. I don't like the populists, but they're maybe a wake-up call to conscientious representatives that they need to do their basic job better.
More garden work
|
Topic: gardening, opinion | Link here |
Mick Solly, the gardener, along today at the crack of dawn (well, 7:50, earlier than I would have liked). He spent 6 hours mowing most of the lawn and tidying up round shrubs. He didn't bring his ride-on lawn mower: his trailer had had two flat tyres. So I lent him mine, and he got about 20 metres before he jammed the cutting blades with wet grass and burnt out the only recently installed drive belt, something that I have never seen before. And we wanted to sell it in the next few days...
He'll be back on Friday with a new belt. And hopefully he'll find time to do the rest of the work.
Two Grevillea rosmarinifolia?
|
Topic: gardening | Link here |
On the corner of Stones Road and Progress Road there are a few Grevillea bushes, Grevillea rosmarinifolia I think. Or I thought. Today it occurred to me that different bushes have somewhat different flowers:
|
|
Are they the same species? Maybe different subspecies? There seem to be a lot of cultivars.
Kebabs again
|
Topic: food and drink, language, opinion | Link here |
Once upon a time there were spits, sticks on which you could put meat for cooking. And the result was also called a spit. But “spit” has negative connotations, so gradually they became skewers. And now, in our multicultural society, most of these skewers have a Middle Eastern touch, so people call them kebabs. I suppose that makes sense, even when making more traditional European styles like the remaining two skewers that we ate tonight.
Last time we cooked the kebabs (there, I've said it) in the air fryer at 230° for 14 minutes, and it was too much. Today, as planned, I reduced it to 10 minutes. Still too much:
|
|
OK, next time 8 minutes.
Moon photos
|
Topic: photography, opinion | Link here |
Full moon today, the one for the Moon cakes festival (now apparently called mid-autumn festival, a singularly appropriate name here). Took a couple of photos:
|
|
Should I have optimized the photos? Without that, the second image would look like the first here (run the cursor over an image to compare it with its neighbour):
|
|
On consideration, I think that it's probably worthwhile.
Sunday, 15 September 2019 | Dereel | Images for 15 September 2019 |
Top of page | ||
previous day | ||
next day | ||
last day |
Bloody Iranians!
|
Topic: politics, opinion | Link here |
One of the more interesting strategies in modern warfare is the use of drones. The Houthis in Yemen have been using them for a while, inflicting some damage on the territory of the Saudi invaders.
But yesterday they damaged key Saudi oil installations. That's a completely different level, and various reports claim (almost certainly incorrectly) that up to half the world's oil supply could be threatened. Did they deny responsibility? No, quite the contrary: they claimed to have done it, which sounds plausible.
But that's all fake news. US Secretary of State Mike Pompeo got up on his hind legs and claimed:
Tehran is behind nearly 100 attacks on Saudi Arabia while Rouhani and Zarif pretend to engage in diplomacy. Amid all the calls for de-escalation, Iran has now launched an unprecedented attack on the world’s energy supply. There is no evidence the attacks came from Yemen.
Fascinating:
I was hoping that after the sacking of John Bolton the USA might become saner again. I can't see any evidence. I'd like to say “I'm disgusted” and be done with it. But this irresponsible behaviour of the USA makes war ever more likely.
Where did my mail go?
|
Topic: technology, opinion | Link here |
For the last couple of days I have had no automatic daily reports from teevee. Why? It's still running normally. Off to take a look.
The first part of the problem was that I had removed the wrong disk when swapping the backup disks I have on eureka. I have two separate photo backup disks, and one is kept at Chris Bahlo's place. But a couple of days ago I accidentally removed the video backup disk, with the result that the nightly backups didn't continue (some hang, strangely). Fix that and run the nightly cleanup script:
=== root@teevee (/dev/pts/10) ~ 27 -> mailme cleanup
It ran, but I got no output. What happened? From /var/log/maillog:
Sep 15 14:30:12 teevee postfix/pickup[70746]: 166EB8ADD6: uid=0 from=<root@teevee.lemis.com>
Sep 15 14:30:12 teevee postfix/cleanup[70791]: 166EB8ADD6: message-id=<20190915043012.166EB8ADD6@teevee.lemis.com>
Sep 15 14:30:12 teevee postfix/qmgr[909]: 166EB8ADD6: from=<root@teevee.lemis.com>, size=21089, nrcpt=1 (queue active)
Sep 15 14:30:12 teevee postfix/smtp[70795]: 166EB8ADD6: to=<root@teevee.lemis.com>, relay=eureka.lemis.com[192.109.197.137]:25, delay=0.22, delays=0.12/0.06/0/0.03, dsn=2.0.0, status=sent (250 2.0.0 Ok: queued as 52B89264C5D)
Sep 15 14:30:12 teevee postfix/smtpd[70796]: connect from eureka[192.109.197.137]
Sep 15 14:30:12 teevee postfix/smtpd[70796]: 7FEC18AF55: client=eureka[192.109.197.137]
Sep 15 14:30:12 teevee postfix/qmgr[909]: 7FEC18AF55: from=<root@teevee.lemis.com>, size=21859, nrcpt=1 (queue active)
Sep 15 14:30:12 teevee postfix/smtp[70795]: 7FEC18AF55: to=<root@teevee.lemis.com>, relay=eureka.lemis.com[192.109.197.137]:25, delay=0.01, delays=0/0/0/0, dsn=2.0.0, status=sent (250 2.0.0 Ok: queued as 93F86264C5D)
...
Sep 15 14:30:13 teevee postfix/qmgr[909]: 5F3E78AF56: from=<>, size=41479, nrcpt=1 (queue active)
Sep 15 14:30:13 teevee postfix/smtp[70795]: 5F3E78AF56: to=<root@teevee.lemis.com>, relay=eureka.lemis.com[192.109.197.137]:25, delay=0.01, delays=0/0/0/0, dsn=5.4.0, status=bounced (host eureka.lemis.com[192.109.197.137] said: 554 5.4.0 Error: too many hops (in reply to end of DATA command))
Sep 15 14:30:13 teevee postfix/qmgr[909]: 5F3E78AF56: removed
A mail loop! But how did that happen? It was working fine before. How did the problem occur? Back looking through /usr/local/etc/postfix/main.cf:
# For the sake of consistency between sender and recipient addresses,
# myorigin also specifies the default domain name that is appended
# to recipient addresses that have no @domain part.
#
#myorigin = $myhostname
myorigin = $mydomain
That looks right. The mail should go to root@lemis.com, not root@teevee.lemis.com.
Oh, wait, that's the From: address. Where does this message get sent? mailme is a shell function:
mailme()
{
cmd=$1
shift
xtset mailme $cmd
date
$cmd $* 2>&1 | tee /dev/tty | mail -s "mailme: $cmd $*" `who am i|awk '{print $1}'`
cd .
date
}
fi
OK, it invokes mail. Try running that. It hangs! Normally mail stops reading input when it gets a line with a single period on it. But today it didn't. It didn't even (directly) respond to Ctrl-C.
Finally it dawned on me:
=== root@teevee (/dev/pts/10) ~ 27 -> wh mail
4256743 -rwxr-xr-x 1 root wheel 212104 4 Jul 12:57 /usr/local/bin/mail
980079 -r-xr-xr-x 3 root wheel 115440 7 Jul 14:49 /usr/bin/mail
And yes, I put /usr/local/bin before /usr/bin in my PATH, which usually works better. But what's that /usr/local/bin/mail doing there? I didn't ask for that.
=== grog@teevee (/dev/pts/5) ~ 76 -> pkg which /usr/local/bin/mail
/usr/local/bin/mail was installed by package mailutils-3.7_1
OK, get rid of that:
=== root@teevee (/dev/pts/10) ~ 21 -> pkg delete mailutils-3.7_1
Checking integrity... done (0 conflicting)
Deinstallation has been requested for the following 2 packages (of 0 packages in the universe):
Installed packages to be REMOVED:
mailutils-3.7_1
emacs-26.2,3
Somewhere I swear there's a conspiracy to ensure that any package removal includes removing Emacs. But clearly that's not an option.
=== root@teevee (/dev/pts/10) ~ 22 -> chmod -x /usr/local/bin/mail
That did the trick. But why did this happen now? After some consideration, it's clearly because I had elevated myself to root with su rather than logging in directly. As a result it used my PATH rather than root's PATH.
Monday, 16 September 2019 | Dereel | Images for 16 September 2019 |
Top of page | ||
previous day | ||
next day | ||
last day |
Spring is here
|
Topic: gardening, opinion | Link here |
Mid-September, a time of month when I used to do my monthly flower photos. I've put that off until the Equinox, but particularly right now the flowers are of interest. The north garden is full of flowering bulbs, and the succulent bed is flowering nicely:
|
|
The Iris are also flowering, including the new dark ones that we got from Petra Gietz:
|
And so are the Alyogyne huegelii:
|
I almost wonder if I want to hide them by planting tomatoes there.
But one plant is giving cause for concern: the Echium that looked so good only last week. Here then (first image) and now:
|
|
|
What has happened? Went looking and discovered something similar for last year.
|
Does it need more fertilizer? Or water? There should be plenty of water.
Web site configuration, next step
|
Topic: technology, opinion | Link here |
I've been dragging my feet on configuring the new web site, but time is moving on. I need to complete the migration by the end of the month, and that includes moving Chris Bahlo's stuff as well. The current status was that Apache works, but I hadn't configured PHP. OK, off to look at the documentation.
What documentation? I didn't find much of use. Though this page gave some ideas, it didn't relate directly to what I was looking for, and it also wasn't official. The official just had an indication to install mod_php71, and at FreeBSD handbook just had an indication to install mod_php56, not exactly the latest release, and said nothing about configuration. I had already done that, and I know there's more to it than that. But I couldn't find any documentation that says what to do, though there was some mention of a file /usr/local/etc/php.ini, which didn't exist.
OK, from memory: you need to install base PHP separately. Why doesn't the module depend on it? And why PHP 5.6? A bit of investigation showed that 7.3 is the latest version, though it's not clear whether it's a supported release or beta. What the hell, install it anyway. That gave me /usr/local/etc/php.ini-production and /usr/local/etc/php.ini-development. Which do I use? What's the difference? It seems mainly related to error reporting.
OK, based on my configuration on the old www, just add the include_path:
--- php.ini 2019/09/07 01:13:29 1.1
+++ php.ini 2019/09/16 05:56:44
@@ -734,7 +734,9 @@
; UNIX: "/path1:/path2"
;include_path = ".:/php/includes"
-;
+
+include_path = ".:/home/grog/www.lemis.com/grog/php/includes"
+
; Windows: "\path1;\path2"
;include_path = ".;c:\php\includes"
;
Is that enough? No, first I need to add stuff to httpd.conf. Also did some comparison with www, and ended up with:
--- httpd.conf 2019/08/22 04:02:47 1.1
+++ httpd.conf 2019/09/16 06:12:48
@@ -178,7 +178,16 @@
#LoadModule userdir_module libexec/apache24/mod_userdir.so
LoadModule alias_module libexec/apache24/mod_alias.so
#LoadModule rewrite_module libexec/apache24/mod_rewrite.so
+LoadModule php7_module libexec/apache24/libphp7.so
+# Does this belong here? Greg 20190916
+<FilesMatch "\.php$">
+ SetHandler application/x-httpd-php
+</FilesMatch>
+<FilesMatch "\.phps$">
+ SetHandler application/x-httpd-php-source
+</FilesMatch>
+
# Third party modules
IncludeOptional etc/apache24/modules.d/[0-9][0-9][0-9]_*.conf
@@ -213,7 +222,7 @@
# e-mailed. This address appears on some server-generated pages, such
# as error documents. e.g. admin@your-domain.com
#
-ServerAdmin you@example.com
+ServerAdmin grog@lemis.com
#
# ServerName gives the name and port that the server uses to identify itself.
@@ -222,7 +231,7 @@
#
# If your host doesn't have a registered DNS name, enter its IP address here.
#
-#ServerName www.example.com:80
+ServerName www.lemis.com:80
#
# Deny access to the entirety of your server's filesystem. You must
@@ -246,8 +255,10 @@
# documents. By default, all requests are taken from this directory, but
# symbolic links and aliases may be used to point to other locations.
#
-DocumentRoot "/usr/local/www/apache24/data"
-<Directory "/usr/local/www/apache24/data">
+# DocumentRoot "/usr/local/www/data"
+# <Directory "/usr/local/www/data">
+DocumentRoot "/home/grog/www.lemis.com"
+<Directory "/home/grog/www.lemis.com">
#
# Possible values for the Options directive are "None", "All",
# or any combination of:
@@ -260,8 +271,8 @@
# http://httpd.apache.org/docs/2.4/mod/core.html#options
# for more information.
#
- Options Indexes FollowSymLinks
-
+ Options FollowSymLinks
+ Options -Indexes
#
# AllowOverride controls what directives may be placed in .htaccess files.
# It can be "All", "None", or any combination of the keywords:
@@ -280,7 +291,7 @@
# is requested.
#
<IfModule dir_module>
- DirectoryIndex index.html
+ DirectoryIndex index.php index.html
</IfModule>
, "1"); /* >>> */ ?>
Doing my comparisons on www, I also found these diffs:
-#LoadModule session_module libexec/apache24/mod_session.so
+LoadModule session_module libexec/apache24/mod_session.so
# Server-pool management (MPM specific)
-#Include etc/apache24/extra/httpd-mpm.conf
+Include etc/apache24/extra/httpd-mpm.conf
What's that? I don't know, but I'm sure I don't need them at the moment, so I put them off until later.
Finally, start. Does it work? Sort of. First I needed to modify /etc/rc.conf:
apache24_enable=YES
Otherwise it won't allow apachectl start: it wanted apachectl onestart, which doesn't exist.
Finally started. It works. But it logged to the wrong place, and in the wrong format; I can check that later. More to the point, /var/log/httpd-error.log grew alarmingly:
-rw-r--r-- 1 root wheel 38,997 16 Sep 07:04 httpd-access.log
-rw-r--r-- 1 root wheel 3,419,426 16 Sep 07:04 httpd-error.log
What's that? The main error message was:
[Mon Sep 16 06:18:11.915844 2019] [php7:notice] [pid 68504] [client 207.46.13.178:5277] PHP Notice: A non well formed numeric value encountered in /home/grog/www.lemis.com/grog/php/includes/onephoto.php on line 1020
[Mon Sep 16 06:18:11.915873 2019] [php7:notice] [pid 68504] [client 207.46.13.178:5277] PHP Notice: A non well formed numeric value encountered in /home/grog/www.lemis.com/grog/php/includes/onephoto.php on line 1080
[Mon Sep 16 06:18:11.915893 2019] [php7:notice] [pid 68504] [client 207.46.13.178:5277] PHP Notice: A non well formed numeric value encountered in /home/grog/www.lemis.com/grog/php/includes/onephoto.php on line 1214
[Mon Sep 16 06:18:11.915916 2019] [php7:notice] [pid 68504] [client 207.46.13.178:5277] PHP Notice: A non well formed numeric value encountered in /home/grog/www.lemis.com/grog/php/includes/onephoto.php on line 1225
That proved to be implicit numerical conversions. It's in the function exposuredetails(), and the variable $focal_length was a string with contents like 30 mm. It got it right, but it didn't like the mm at the end. OK, that's easy enough to fix.
And then I got other messages:
[Mon Sep 16 06:19:50.920874 2019] [php7:notice] [pid 68505] [client 207.46.13.103:3503] PHP Notice: Uninitialized string offset: 0 in /home/grog/www.lemis.com/grog/php/includes/utils.php on line 29
That's also fixable, apparently a parameter passing error that I can check somewhere. But where's that coming from? What's 207.46.13.103? httpd should have resolved the name before logging, but that's on the to-do list. It proves to be msnbot-207-46-13-103.search.msn.com.. I've only just brought up this system (w3.lemis.com) and already the search bots are crawling it. OK, shut down again, lick my (few) wounds, and plan what to do next:
And when it's all done: document the procedure for people who come after.
Tuesday, 17 September 2019 | Dereel | Images for 17 September 2019 |
Top of page | ||
previous day | ||
next day | ||
last day |
Power failures again
|
Topic: Stones Road house, general | Link here |
Another two short power failures today. Like last week they were between 12:00 and 13:00. Are they doing some kind of work during that time?
NBN not to be beaten
|
Topic: technology, general, opinion | Link here |
Two power failures? 10 seconds? Hah! The National Broadband Network can do better. Today we had 6 failures (sorry, “scheduled maintenance”) with a total of 7058 seconds, nearly 2 hours. That's worse than it seems: somehow it completely disrupts the day. When writing this diary, for example, I frequently check things on the web, something that kept failing on me. As a result, I didn't get the diary committed until 16:24.
How I hate the NBN! Why can't they get their act together?
Range hood issues
|
Topic: general, Stones Road house, opinion | Link here |
I've complained at length about the underpowered range hood, starting almost immediately after moving in here. My first reference in this diary was on 26 May 2015, but it referred to complaints I had made earlier.
After the disastrous results of my assessment from the DBDRV, I've given up on expecting correction of the deficiencies in the house, including the range hood. Clearly it works up to a point: a week or so ago I found drops of oil in the central pane of mesh that passes for a filter, so I put it through the dishwasher. But today I found more oil. Where did that come from? Looking upwards after removing the pane I saw:
|
Much of the oil has condensed on the motor. Yes, I can wipe it off, but what does that do for the lifespan?
Wednesday, 18 September 2019 | Dereel | Images for 18 September 2019 |
Top of page | ||
previous day | ||
next day | ||
last day |
Powercor: outdoing NBN?
|
Topic: general, opinion | Link here |
Still more power failures today, once again round midday, once again relatively short, but long enough to cause problems without the photovoltaic system: three of them with duration between 4 and 11 seconds. What are these people doing? That's now been 33 power failures since we installed the system, but they've been very unevenly spaced: none at all in the first month, and 11 in the last week. The good news is that none of them have had any effect on our power.
Could it be the inverter getting confused? Potentially, but the fact that it happens at pretty much the same time of day suggests that there's something else at play.
Schinkengriller!
|
Topic: food and drink, opinion | Link here |
The Oktoberfest is on us again. Or at least, at ALDI. Yvonne off today, bringing back lots of cheap beer and a few other items, including „Schinkengriller“ sausages (a word which means “ham griller”, in other words a sausage designed for grilling).
Tried some this evening in the “hair dryer” (the bigger of the two “air fryers”). They all burst:
|
|
Would that happen if they were grilled? I almost suspect that it would.
White balance pain
|
Topic: photography, opinion | Link here |
The photos of the „Schinkengriller“ were interesting not just because of the subject material, but because of the colour balance:
|
|
They were both taken with the OM-D E-M1 Mark II, the M.Zuiko Digital ED 12-100 mm f/4.0 IS PRO and the mecablitz 15 MS-1 ring flash. Particularly the flash should have ensured that the colour balance was correct. Did the camera miscalculate? No, it's not that simple. I tried setting the white balance for both photos to be the same, but they looked completely different. Changing the white balance to come close to the same sausage colour made the background look silly before it came close to setting the same colour for the sausages.
As if that wasn't enough, Ashampoo optimizer didn't do a good job either, adding a significant greenish tinge that I've seen before (run the cursor over an image to compare it with its neighbour):
|
|
I suppose the reddish colour of the sausages confused it, but it doesn't explain the difference in the original images.
Thursday, 19 September 2019 | Dereel | Images for 19 September 2019 |
Top of page | ||
previous day | ||
next day | ||
last day |
Web site update, one more small step
|
Topic: technology, opinion | Link here |
I really need to get round to completing my web site migration, but somehow I never find time for it. In the meantime it occurred to me that I'm not alone on the site: Chris Bahlo also has narrawin.com on the same machine, so we'll need to move that too. Moved her stuff across and sent her a mail message, cleverly designed to arrive when she's at a mediaeval tournament somewhere in St Ives, so she won't have any time to look at it.
Then took a look at my PHP problem. This code:
if ($alignments [0])
{
switch ($alignments [0])
...
gave rise to this error message:
[Mon Sep 16 07:00:54.671353 2019] [php7:notice] [pid 68500] [client 85.208.96.7:56297] PHP Notice: Uninitialized string offset: 0 in /home/grog/www.lemis.com/grog/php/includes/utils.php on line 29
OK, follow what's going on. Did I not pass a correct string? After some searching discovered yes, I did, but the code was in a loop:
foreach ($cols as $col)
{
switch ($alignments [0])
...
$alignments = substr ($alignments, 1); /* shift left */
}
That works fine in C, but PHP doesn't like people looking at empty strings, thus the message. Once I understood that, fixing it was simple:
if (strlen ($alignments) > 1)
$alignments = substr ($alignments, 1); /* shift left */
Next, what are these strange modules that I saw on Monday. What are session_module and MPM? Something that Chris wanted, maybe? Off looking in my diary and found this entry, which didn't really make things that much clearer. But at least now I have a lead.
And that's about all I did, though I'm now down to only two immediate things to do:
More kangaroos
|
Topic: animals, general | Link here |
For some obscure reason the kangaroos are out in force again:
|
|
It was very windy today, but I can't see any reason why that would make a difference.
Another power failure
|
Topic: Stones Road house, general | Link here |
Yet another power failure today, similar to the ones I've been having all week. This was the shortest, at only 2 seconds, and there was only one.
What are they doing?
Shut up, bloody chrome!
|
Topic: technology, opinion | Link here |
Somehow I have had problems with web browsers almost since they came into existence. Each new browser seems to be worse than the one that came before it. mosaic, then netscape, then firefox. firefox is clearly on the way out, something that you'd think I would welcome. But I kept it because I was able (with some pain) to remap the keys to something closer to Emacs.
How could that be allowed to happen? So firefox broke the interface that allowed that kind of addon. No advantage to firefox any more, and at least one significant disadvantage:
PID USERNAME THR PRI NICE SIZE RES STATE C TIME WCPU COMMAND
5370 grog 89 20 0 4917M 3390M uwait 6 18.6H 14.79% firefox
5330 grog 53 20 0 1090M 514M uwait 7 744:53 4.20% firefox
32079 grog 51 20 0 2007M 1276M uwait 6 145:28 2.10% firefox
93996 grog 56 20 0 1369M 857M uwait 6 10:31 0.10% firefox
5045 grog 4 20 0 704M 122M select 5 4:20 0.20% emacs-24.5
Look at that memory usage! Four instances using a total of 9,393 MB of address space and 6,037 MB of RAM! mosaic got by on an 8 MB system! And doesn't last century's memory hog, Emacs, look insignificant by comparison! I'm really amazed how a browser can use so much memory. Yes, they keep adding things to them, but surely you could do that in a tenth of the space. Each instance seems to swallow 1 GB of memory before it even gets started.
Now that firefox has removed its one advantage over chrome, why not use that instead? It has some positively horrible paradigms, but then so has firefox. So today I fired one up, marveling at the way it positioned itself a little too far down on the screen, so that it ran off at the bottom. Maybe that's modern, but my guess is that it has miscalculated window frame widths (something that it wants to do itself to make it incompatible with everything else on the screen). But it worked as far as I tried.
But then in the evening, while watching TV, Chrome suddenly took over the screen. “Can't update Chrome”. That's OK, I don't want you to. But what insolence on the part of the developers to interrupt what I'm doing for an error message about a function I neither asked for nor want!
Friday, 20 September 2019 | Dereel | Images for 20 September 2019 |
Top of page | ||
previous day | ||
next day | ||
last day |
More garden work
|
Topic: gardening | Link here |
Mike Solley, the gardener, along this morning to repair the damage he did to the lawn mower last week, and to continue preparations for spring. Things are looking a lot tidier now.
Selling the lawn mower
|
Topic: gardening, general, opinion | Link here |
A Jason along to look at the lawn mower today, bringing a trailer with him. He looked very interested, and tried a token “beat you down $100”, which I turned down. There will be others.
But Yvonne didn't like that. I should have taken the $1,100 and run. Dammit, domestic peace is worth more than $100. She contacted Jason, and he agreed to pick it up on Sunday.
Modern web language
|
Topic: language, music, opinion | Link here |
Yvonne didn't know why the Great C Major was called that: there's also a Little C Major. What was the number again? Off to ask Google. One of the high-ranking answers [now mercifully dead] blew my mind:
Symphonie Nr. 6 C-Dur D 589 "Kleine C-Dur" - Song By Wolfgang ...Explore Symphonie Nr. 6 C-Dur D 589 "Kleine C-Dur", a song first recorded by Wolfgang Amadeus Mozart. Where did they get Mozart from?View Symphonie Nr. 6 C-Dur D 589 "Kleine C-Dur" ...
How many things are wrong about this!
To be fair, this was only Google's representation of the page. I didn't find the text in the (almost empty) entry, and at least they now claim that it was Schubert who “recorded” the “song”. But this kind of stupidity turns my stomach.
Saturday, 21 September 2019 | Dereel | Images for 21 September 2019 |
Top of page | ||
previous day | ||
next day | ||
last day |
Indian mee goreng?
|
Topic: food and drink, opinion | Link here |
For some time I've been considering trying a dish in Wendy Hutton's “Singapore Food”: “Indian Mee Goreng”. The name belies its international nature: “Mee” (now “Mi” in Malay, “Mie” in Indonesian, pronunciation unchanged) is Straits Chinese for yellow noodles, “goreng” is Malay for “fry” or (in this case) “fried”. And despite the name, noodles are not part of the Indian cuisines.
This one interested me because it included karuveppilai (Curry tree leaves). But how authentic is it? Out looking for more recipes, coming up with some interesting results. Some of the recipes were called “Mee goreng mamak”. What's Mamak? It refers to Tamil immigrants who have embraced the Muslim faith, and they're apparently the originators of this mix.
There's this recipe from Kampung Singapura, which doesn't look bad, and this one from Nyonya Cooking, and a couple that I got rid of because of their insistence on irritating popups.
A number linked back to this recipe in Rasa Malaysia. This one is interesting because, though written in English, it includes ingredient names in Chinese (but not in Tamil):
It also comes up with the name “Uncle's fried noodles”. Why uncle? It seems that that's the original meaning of “mamak”, though it's not clear in what language. It sounds Malay, but it's not in my dictionary. It doesn't sound Tamil or Chinese. Wikipedia doesn't (yet) have an entry for it, though Mamak stall explains:
The Malaysian Mamak are Malaysians of Tamil Muslim origin, whose forefathers mostly migrated from South India to the Malay Peninsula and various locations in Southeast Asia centuries ago.The word 'Mamak' is from the Tamil term for maternal uncle, or 'maa-ma'. In Singapore and Malaysia, it is used by children as an honorific to respectfully address adults such as shopkeepers. The silent K in 'Mamak' likely came about as a hypercorrection; since terminal Ks are not pronounced in Malay, a Malay who heard the Tamil word may have assumed there was a silent K at the end.
The things you learn.
Preparing Indian mee goreng
|
Topic: food and drink, opinion | Link here |
In the end I started with the Wendy Hutton recipe for Indian Mee Goreng, and modified it based on some of the other recipes:
quantity | ingredient | step | ||
450 g | Hokkien noodles | 1 | ||
90 g | squid tubes | 2 | ||
90 g | firm dofu | 2 | ||
110 g | fresh tomato | 2 | ||
160 g | cooked potato | 2 | ||
90 g | onion | 2 | ||
95 g | soya sauce | 3 | ||
38 g | dark soya sauce | 3 | ||
38 g | sambal ulek | 3 | ||
38 g | tomato sauce | 3 | ||
1 | egg | 4 | ||
oil | 5 | |||
100 g | taugeh | 7 | ||
And the curry leaves? I forgot them! Next time. It didn't taste bad, though I somewhat overdid the dark soya sauce.
|
It's quite heavy, and I don't see myself eating it every day.
Installing the new sprinkler relay
|
Topic: technology, gardening, opinion | Link here |
The weather's getting warmer, and I still haven't installed my new sprinkler relay. How did that work? Good thing I documented it when I received it. OK: first thing is that it still has an RFC 1918 address. Add an alias:
=== root@eureka (/dev/pts/11) ~ 80 -> ifconfig em0 192.168.123.123/16
Aargh! Too late I noticed that I had forgotten the all-important alias! OK, I can reset it:
=== root@eureka (/dev/pts/11) ~ 81 -> ifconfig em0 eureka
^C
Damn! I can't access any useful name server. OK, use the IP address:
=== root@eureka (/dev/pts/11) ~ 82 -> ifconfig em0 192.109.197.137
=== root@eureka (/dev/pts/11) ~ 83 -> ping teevee
ping: cannot resolve teevee: Host name lookup failure=== root@eureka (/dev/pts/11) ~ 85 -> ping 192.109.197.134
PING 192.109.197.134 (192.109.197.134): 56 data bytes
64 bytes from 192.109.197.134: icmp_seq=0 ttl=64 time=0.352 ms
^C
OK, we're on the LAN, but named has got itself into a huff and isn't responding. It should be running locally, right?
=== root@eureka (/dev/pts/11) ~ 87 -> cat /etc/resolv.conf
search lemis.com
nameserver 192.109.197.137
OK, let's SIGHUP it:
=== root@eureka (/dev/pts/11) ~ 88 -> ps aux|grep named
bind 1747 0.0 0.1 100444 24752 - Is 11Sep19 2:11.41 /usr/local/sbin/named -u bind -c /usr/local/etc/namedb/named.conf
root 93423 0.0 0.0 18848 2320 11 S+ 2:47pm 0:00.00 grep named=== root@eureka (/dev/pts/11) ~ 89 -> kill -1 1747
=== root@eureka (/dev/pts/11) ~ 90 -> ps aux|grep named
bind 1747 0.1 0.1 100444 34352 - Ss 11Sep19 2:11.45 /usr/local/sbin/named -u bind -c /usr/local/etc/namedb/named.conf
root 93932 0.0 0.0 18848 2320 11 S+ 2:48pm 0:00.00 grep named=== root@eureka (/dev/pts/11) ~ 91 -> ping lagoon
ping: cannot resolve lagoon: Host name lookup failure
Still no reaction. Dammit, let's restart it the normal way:
=== root@eureka (/dev/pts/11) ~ 95 -> service named restart
named not running? (check /var/run/named/pid).
named already running? (pid=1747).
How I hate these silly error messages framed as questions! The situation doesn't help. Let's go back to the traditional way, at least in part:
=== root@eureka (/dev/pts/11) ~ 96 -> killall named
=== root@eureka (/dev/pts/11) ~ 97 -> service named start
Starting named.=== root@eureka (/dev/pts/11) ~ 99 -> ping lagoon
PING lagoon.lemis.com (192.109.197.134): 56 data bytes
64 bytes from 192.109.197.134: icmp_seq=0 ttl=64 time=0.288 ms
Finally! And that wasn't even the job I wanted to do. Back to the relay board. Yes, after adding the alias, I was able to talk to it and set the IP address and port (sprinkler.lemis.com, 192.109.197.180, port 80).
Ah, but that's an old, worn-out meaning of the word “port”. What the modern programmers of this board mean is a web page name. By default it was http://192.168.1.4/30000. Now it's http://sprinkler.lemis.com/80. OK, I can live with that, especially since it worked first time. I still don't know how to reset it, but I'll leave that to another day.
And next? I have to simulate web requests, because that's the only way I know to talk to the box. That should be simple enough, but how about checking that I can talk to it in situ first.
No! That's strange: I had been out there with euroa, my Microsoft laptop, and I was able to talk to the network with that. But it took a while, and it occurred to me that it had probably switched to the 802.11 interface.
So: the problem is probably not the old relay board (saves programming), but the network cabling between the pantry and the shed. So I have a completely different issue to solve.
Debugging network cabling
|
Topic: technology, opinion | Link here |
Where do I start with the network cabling? In the middle, I would think. The house network is centred on top of a fridge in the pantry, behind a microwave oven that runs on the same wavelength:
|
That's what it looks like now; when I installed it we had:
|
The additional plate on the right is for teevee, installed when we moved the TV into the lounge room. So we have a total of 9 network connections: 8 rooms in the house, teevee and sprinkler. That's a large value of 9. Which is the connection for sprinkler? Where's my list?
I couldn't find it. But one thing's relatively clear: there's an issue with the cable, and just locating it doesn't mean it's fixed. What alternatives do we have?
Alternative network link
|
Topic: technology, opinion | Link here |
What do you do when you don't have an Ethernet cable? Wi-Fi, that silly name for 802.11. And how about that, I have a second Wi-Fi access point. Why not put that in the shed and run it as a bridge to the access point in the pantry?
One reason, I discovered, is because it's very hard to configure, at least with these el-cheapo boxes. The good news is: yes, they offer the function. The bad news is: so far they haven't delivered.
There are clearly some issues with the web interface: they don't show the client list, for example, although taskumatti was connected to the pantry AP (air-gw-1) the whole time. And the constant requirement to log in to the things makes me wish I had chosen simpler passwords.
The first issue was the rendering of the web page:
|
That was on firefox; here Chrome was better:
|
There was a reasonable amount of help on the page:
Enable Wireless RepeatingEnable this if you wish to use either Bridge mode or Repeater mode, and then select the mode you want for your environment.
Disable Wireless Clients Association
With this feature enabled, the client will not be able to access the LAN
Wireless Repeater.
In this mode, the "+host_name+"router will communicate only with another Base Station-mode wireless station. You must enter the MAC address (physical address) of the other Base Station-mode wireless station in the field provided. WEP can (and should) be used to protect this communication.
Wireless Base Station.
Select this only if this "+host_name+"router is the master for a group of Repeater-mode wireless stations. The other Repeater-mode wireless stations must be set to Wireless Repeater mode, using this "+host_name+"router`s MAC address. They then send all traffic to this master, rather than communicate directly with each other. WEP can (and should) be used to protect this traffic.
If this option is selected, you must enter the MAC addresses of the other access points in the fields provided.
Clearly the "+host_name+" was intended to be replaced by the marketing name of the router (Goldweb, Copyright © XXX Systems, Inc., according to the web pages). So I set things as described, first tripping over this detail:
|
That's the slave side. The IP address is the address of this (slave, repeater) AP, the MAC address is the address of the other (“Basic station”) AP. First time round I set the IP of the base station, which it happily set, taking both off the air in the process. And the “LAN Configuration” refused to set the address for some reason; I could only set it here.
In the end I had the confirmation on base station and repeater respectively:
|
|
So it seems that it should work. But it didn't. I was able to associate with the slave AP (by disabling association with the master/base station), but it only worked when the slave was connected to eureka with an Ethernet cable. Do I have issues with overlapping IP addresses? That shouldn't be a problem. Or ARP on the other side? This stuff is far too painful.
PV inverter calibrating again!
|
Topic: Stones Road house, general, technology, opinion | Link here |
My solar electricity inverter has started calibrating batteries again, running 18 hours from 2019-09-20 17:53:43 until 2019-09-21 11:54:43, and once again draining battery capacity to 0. It's been a month since the last time, and on that occasion Fred had promised to get a statement from the manufacturers of both the inverter and the battery. Nothing yet. Time to set a deadline: 11 October, which should give them more than enough time. After all, next month they should be installing the other battery. If they're not compatible, they'll have to install an alternative instead.
Sunday, 22 September 2019 | Dereel | Images for 22 September 2019 |
Top of page | ||
previous day | ||
next day | ||
last day |
More network pain
|
Topic: technology | Link here |
While looking for something completely different today, saw this in the log files, repeating about once a minute:
Sep 22 09:04:11 eureka grog: Restarting SMTP tunnel
Sep 22 09:05:31 eureka grog: Restarting SMTP tunnel
That's the tunnel that enables me to send mail from my local network to the world, bypassing Aussie Broadband's well-meant blocking of outgoing port 25. But why is it failing? Try restarting manually:
=== grog@eureka (/dev/pts/31) ~ 1 -> mailtunnel
ssh: connect to host mail.lemis.com port 22: Host is down
Damn! That's www.lemis.com. Has it crashed?
=== grog@eureka (/dev/pts/31) ~ 3 -> ping www
PING www.lemis.com (208.86.226.86): 56 data bytes
64 bytes from 208.86.226.86: icmp_seq=0 ttl=49 time=262.219 ms
...
No, that's OK. So what's going on? Look at the message again. It's looking for mail.lemis.com, not www. The address is 192.109.197.81, but it's really an alias at www. Why? I'm sure I once had a good reason, maybe to ensure that it's in my /24.
OK, has it forgotten the alias? ifconfig tells me:
xn0: flags=8843<UP,BROADCAST,RUNNING,SIMPLEX,MULTICAST> metric 0 mtu 1500
options=503<RXCSUM,TXCSUM,TSO4,LRO>
ether 00:16:3e:06:34:53
inet 208.86.226.86 netmask 0xfffffffc broadcast 208.86.226.87
inet6 fe80::216:3eff:fe06:3453%xn0 prefixlen 64 scopeid 0x2
inet 192.109.197.81 netmask 0xffffff00 broadcast 192.109.197.255
nd6 options=29<PERFORMNUD,IFDISABLED,AUTO_LINKLOCAL>
media: Ethernet manual
status: active
No, it's there. Damn! Have RootBSD forgotten the routing again? Where does the traceroute end?
=== root@eureka (/dev/pts/11) ~ 123 -> traceroute mail
traceroute to mail.lemis.com (192.109.197.81), 64 hops max, 40 byte packets
1 *^C
Nowhere! Bing! This is (almost) the same address as in my local network. And when I recovered the mess I made yesterday, I set the net mask to 0xffffff00:
em0: flags=8843<UP,BROADCAST,RUNNING,SIMPLEX,MULTICAST> metric 0 mtu 1500
options=4019b<RXCSUM,TXCSUM,VLAN_MTU,VLAN_HWTAGGING,VLAN_HWCSUM,TSO4,VLAN_HWTSO>
ether bc:5f:f4:c9:9b:bf
inet6 fe80::be5f:f4ff:fec9:9bbf%em0 prefixlen 64 tentative scopeid 0x1
inet 192.109.197.137 netmask 0xffffff00 broadcast 192.109.197.255
inet 192.168.123.123 netmask 0xffff0000 broadcast 192.168.255.255
nd6 options=29<PERFORMNUD,IFDISABLED,AUTO_LINKLOCAL>
media: Ethernet autoselect (1000baseT <full-duplex>)
status: active
That meant that the system looked for mail locally rather than go out to the global Internet. Change it:
inet 192.109.197.137 netmask 0xffffff80 broadcast 192.109.197.255
After that, all worked normally again.
Another bloody PV recalibration!
|
Topic: Stones Road house, technology, general, opinion | Link here |
Nice, bright sunshine today, but somehow the PV system didn't want to know. Checking, found that it had entered another recalibration cycle. Damn that. I'll power cycle the thing.
That worked, though it takes the inverter a while to become fully functional again. In fact, it took a very long while. It worked fine, but I couldn't access it on the network. What happened? Looking through the output of arp, I saw:
? (192.109.197.201) at 00:25:ca:0b:fe:05 on em0 expires in 1140 seconds [ethernet]
Damn, I should really write down the MAC addresses of my devices. But then, that's what dhcpd does, in /var/db/dhcpd/dhcpd.leases. And sure enough:
lease 192.109.197.246 {
starts 0 2019/09/22 02:21:20;
ends 0 2019/09/22 02:28:31;
tstp 0 2019/09/22 02:28:31;
cltt 0 2019/09/22 02:21:20;
binding state free;
hardware ethernet 00:25:ca:0b:fe:05;
uid "\001\000%\312\013\376\005";
}
Same MAC address. Why did it change? I thought you could rely on DHCP servers to keep the same IP address as long as there were no conflicts. And where's the entry for the new IP address? Never mind, I have other fish to fry. Update my DNS config to have inverter.lemis.com point to 192.109.197.201, and it worked again.
Network problems, part 2
|
Topic: technology, opinion | Link here |
Just to annoy me, my mobile phone, taskumatti.lemis.com, beeped at me again: “beep-beep....beep-beep” What does that mean? It shows a dark grey on black message for about 5 seconds, and then it goes away, never to be seen again: when I turn the thing on, there's no sign that it ever existed.
OK, in the past it has meant “we're off the net”. Took a look at the status. Yes, off the net.
Why? The rest of the network was working normally. Look more carefully at the configuration: IP address 192.109.197.200, name server 192.109.197.152, network gateway 192.109.197.152.
That's wrong! The name server and gateway are eureka, 192.109.197.137, and taskumatti should be 192.109.197.239. But 192.109.197.152 rings a bell: that's air-gw-1.lemis.com, the master Wi-Fi AP. Checked the configuration: yes, it has taken it upon itself to be a DHCP server, something that I most certainly did not ask for.
OK, reset that, and all was well. But 192.109.197.200 also rings a bell: the new address for inverter.lemis.com was 192.109.197.201, and a check for the range for eureka's dhcpd shows that it starts at 192.109.197.224. OK, that explains the problems I had with the inverter. It also means that the inverter has no longer any connection with the global Internet. I'll have to wait until the lease expires, after 24 hours.
Tracing the network fault
|
Topic: Stones Road house, technology, gardening, opinion | Link here |
All that was noise keeping me from my real business of the day: finding out what's wrong with the house network cabling. Why do I have 9 connections in the pantry for 10 network connectors? What matches what?
Off with a real laptop (eucla.lemis.com, an ancient Dell Inspiron 5100 running FreeBSD 9.0) to take a look. The results were not encouraging. I didn't get round to checking the hard-to-access jack behind my bed, but I found at least two jacks that were dead: in the guest room and the second one in the lounge room. By the end I had established the following mapping in the pantry:
Wall plate | Main switch | AP | Location | |||
1 | 1 | Dining room | ||||
2 | 2 | Greg office | ||||
3 | 3 | Yvonne office | ||||
4 | 4 | Greg bedroom (?) | ||||
5 | 5 | library | ||||
6 | 1 | Yvonne bedroom | ||||
7 | 3 | dead (guest room?) | ||||
8 | 2 | dead (lounge room?) | ||||
9 | 8 | lounge room | ||||
7 | 4 | interconnect | ||||
Connection 7 was not plugged in; I can't recall why, but it didn't work. It's unlikely that that was the connection to the shed. And wall plate connection 8? It could have been the shed, but that doesn't explain the dead connector in the lounge room.
Still, there are at least 2 dead connections. How can I tell whether the one in the shed is one of them? I did get as far as confirming that it didn't work at all with eucla, so it's not clear how much searching will help. About the best I can guess is that this is yet another legacy of the Great Jim Lannen.
So where do I go from here? Bridging my APs didn't work. How about looking for dedicated Ethernet/Wi-Fi bridges? Off to eBay, where I found a large number of devices with dubious descriptions, mainly from sellers who sell a choice of cheap items or the real thing to get themselves higher in an order-by-price list.
But one description referred to powerline Ethernet. Bingo! I bought a couple of TP-Link TL-PA411 power line adapters 6 years ago. They were pretty terrible performance-wise, but for this purpose they'd more than fit the bill. Where are they? Off to look for them, so far without success.
Wildflowers in Grassy Gully Road
|
Topic: gardening, opinion | Link here |
While walking the dogs down Grassy Gully Road, came across this plant:
|
|
I've seen things like that before down Misery Creek Road, but not so close to home.
Also a freshly growing Xanthorrhoea:
|
|
It's interesting to note the smooth, shiny surface of the shoot. I wonder why they're so often bent, like this one.
Goodbye lawnmower
|
Topic: general, gardening | Link here |
Jason along today as promised to pick up the lawn mower. I think I'm happier about the space in the shed than the money he paid for it.
He didn't have any rails to drive the thing onto the back of his trailer, so we had to lift it. To my surprise we managed it together.
Monday, 23 September 2019 | Dereel | Images for 23 September 2019 |
Top of page | ||
previous day | ||
next day | ||
last day |
Garden flowers at vernal equinox
|
Topic: gardening, opinion | Link here |
Today was the Vernal Equinox. I've decided to take my monthly garden photos aligned with the equinoxes and solstices, so today was the day.
There's good news and bad news. The good news is that the bulbs in the north bed and the succulents nearby are all flowering better than ever before. Here the north bed this time last year and now:
|
|
The succulent bed was so insignificant last year that I didn't even take a photo of it. Here it is this year:
|
|
And the plants in the north-west bed that we planted last year have transformed it almost beyond recognition:
|
|
What's that enormous bush at front right? I thought it would flower, but there's no evidence yet:
|
It will have to prove itself or leave.
And the ill-fated Clematis “General Sikorski” is still not dead.
|
Clearly it's sensitive to other plants invading its root space.
On the other hand, various plants have suffered or probably even died. Here the Echium and the Euphorbia, last year then this year:
|
|
|
|
And here the Buddlejas in the south:
|
|
Last month's lonely climbing “Iceberg” rose has not stopped flowering, but now the blooms aren't as lonely:
|
|
|
Waterproof?
|
Topic: photography, opinion | Link here |
Walking the dogs, we got caught in a rain shower, even with hail. My camera? No worries, it's the Olympus OM-D E-M1 Mark II with the M.Zuiko Digital ED 12-100 mm f/4.0 IS PRO, both water resistant.
But taking the garden photos photos later on, I discovered:
|
What's that haze? It was visible through the viewfinder, but at first I thought that it was my eyes. On closer examination it proved to be moisture between the front element of the lens and the protective filter. That's no disproof of the moisture resistance of the camera and lens, but certainly an indication of the limitations. Here before and after changing to the M.Zuiko Digital ED 14-150 mm f/4.0-5.6:
|
|
More network pain
|
Topic: Stones Road house, technology, opinion | Link here |
So how do I fix the problems with my network link to the shed? The correct solution is to fix the cable connection, but so far I haven't even been able to establish where the other end is. Is it maybe the connector?
|
A few cobwebs, yes, but the contacts look fine.
OK, I've tried the next step: use my second Wi-Fi AP to bridge the network connection. But though it appears to be correctly configured, it doesn't work. Potentially it isn't designed to bridge Ethernet inputs, only Wi-Fi cards.
Next, the powerline Ethernet adapters. But where are they? Out to the shed to look for stuff, and indeed found lots of things: many old mains power adapters, two ancient laptops, an old Ethernet switch and an ADSL gateway box. But not the powerline Ethernet. Have I given them away, maybe?
So at the end of the day I still don't know how to solve it.
Insurance claims, modern style
|
Topic: general, opinion | Link here |
Call from mumble today, who proved to be from a company that calls itself either Resilium or Non Fleet Recoveries. She first—once again!—asked me to identify myself, and was very surprised when I asked her to identify herself. “I'm from {Resilium|Non Fleet Recoveries}!”. But she was able to identify herself by the time of the email message that I had sent (though she omitted the seconds, either because they're she doesn't consider them to be important, or because her toy MUA didn't show them).
She wanted to talk about the accident on 1 March 2019. Sorry, I wasn't involved in any accident on that date. She said she had some questions, but first I still had to identify myself. In the end I told her the location of the accident and the make of my car. She had really wanted the registration. What does that help? The claimant knows all that, and since he supplied the phone number, there's no reason why it couldn't all be fake.
OK, it's all in my email, I said. Yes, she had read it twice:
> Based on the information provided to us, we have determined that you
> are responsible for the loss incurred by our insured.
What information is this? To the best of my knowledge, you have no
information beyond what your insured told you, and he was not present
at the accident. Based on his actions, I have cause to doubt his good
will.
You don't include the repair quote, but I have received one in the
past showing damage to the rear of a car which is inconsistent with
any damage I might have caused to the vehicle.
To summarize: while backing out of a parking space, I grazed a car
behind me with the towbar of my Holden Commodore VZ station wagon. I
got out and saw a scratch on the rear apron of the car, and I assumed
that I had caused it. Looking at the photos of the rear of the car,
however, I find that there were at least three separate scratches, up
to a metre apart and none of them consistent with the way I touched
the car. I cannot accept responsibility for more than one of these
scratches, and I have my severe doubts whether I caused any of them.
If you have other information, please let me know.
Her first question was: “Do you dispute the claim?”. Yes, of course. I read the relevant part of the email to her. Things went round in circles for a while, and then she asked me “did you receive the photos?”. So I read the relevant part of the email to her. “Did you receive all the photos?”. How could I possibly know? But how can this damage be consistent with a single impact with this tow bar?
|
|
|
|
|
|
Arguably the angle could be explained by the suspension, but I really can't see how it could have caused more than a single scratch, and there's nothing that would explain the damage at the side. Even if I had hit the car with two different parts of my car, it wouldn't cause a scratch of that depth, and it would also have left a mark on my car.
In the end I asked to be connected to her supervisor. Line went dead, so I hung up. When she called back, she had not contacted her supervisor, but had new information. OK, I'll talk to you until I get too annoyed. “Firstly, can I confirm your identity?”. OK, I'm too annoyed. Connect me to your supervisor. Oh, and could you please spell mumble? “It's on the mail I sent you”. No, it wasn't: they were too polite to send the message as a mail body, preferring to do it in PDF instead. So she spelt it out: “Arpana”. I thought that that was a Government body, but that proves to be spelt differently.
Later I checked the PDF: signed “Amol”. What's wrong with these people?
Supervisor was out to lunch, but would call me back. She did so while we were walking the dogs, but during that time I had decided that I didn't want to talk to anybody on the phone any more. Do it in writing. That way we have evidence of the nonsense.
Yet another recalibration!
|
Topic: Stones Road house, technology, opinion | Link here |
I've taken to monitoring my PV generation more carefully, with a display on one of the monitors. And today I saw that it had started yet another recalibration, from 12:28:23 to 13:51:27. I decided to leave it, but this time things went differently: it charged from the grid, although there was enough sunlight, but it didn't discharge:
|
The pinkish line is PV power, and the purple line is grid power. Though there was plenty of sun, the inverter decided to charge from the grid, and on hitting 99% (apparently) battery charge it did mainly a discharge, though it did use some PV power as well. What does that all mean?
Tuesday, 24 September 2019 | Dereel → Ballarat → Dereel | Images for 24 September 2019 |
Top of page | ||
previous day | ||
next day | ||
last day |
Insurers reply
|
Topic: general, opinion | Link here |
Got a reply from Resilium or Non Fleet Recoveries today, carefully ignoring my questions, and claiming that the damage to the insured's car were consistent with being hit in the manner I describe. But they offered a discount—why? And I can get an independent assessor to look at the photos. Clearly they're not as confident as the barking suggests.
Didn't do anything about it today. This requires preparation, and I had other things to do.
Sprinkler network connection: enough!
|
Topic: Stones Road house, general, technology, opinion | Link here |
I've been trying to work around the network problems that I'm having with my sprinkler system for two days now. I haven't exhausted all possibilities: if I can find the other end of the network cable, I can test it and possibly fix the fault. If I can find a way for my access points to bridge as I want them to, I can use them. Or I could find those powerline Ethernet adapters.
But I'm exhausted. I've had enough. Powerline adapters are available on eBay for round $35 a pair, delivery next week. Officeworks has them for $80, but today.
That's enough. Into Ballarat to Officeworks and picked up a couple of T-Link AV600 units. Back home and set them up. They have three LEDs, conveniently marked with intuitive icons:
|
|
From top to bottom, they seem to mean:
They don't seem to have improved the shape since I bought my last ones. They're still too wide for a standard Australian socket:
|
OK, plugged one in in the kitchen and the other in the dining room. Top two LEDs on, as expected. Took the kitchen one to the shed, again top two. Connected to the relay board. Third LED went on. Great!
But then the middle light went out. No network. Damn! Back to the dining room. All lights out. Why is that? Contact problems because of the stupid shape?
Put it in the office behind my desk, and all seemed well:
=== grog@eureka (/dev/pts/33) ~ 16 -> ping sprinkler
PING sprinkler.lemis.com (192.109.197.180): 56 data bytes
64 bytes from 192.109.197.180: icmp_seq=0 ttl=128 time=9.505 ms
64 bytes from 192.109.197.180: icmp_seq=1 ttl=128 time=6.772 ms
64 bytes from 192.109.197.180: icmp_seq=2 ttl=128 time=8.118 ms
That's not exactly the kind of ping time you would expect from a device that promises to be 600 Mb/s. By contrast, my 10 Mb/s office phone ATA has:
=== grog@eureka (/dev/pts/33) ~ 17 -> ping officephone
PING officephone.lemis.com (192.109.197.226): 56 data bytes
64 bytes from 192.109.197.226: icmp_seq=0 ttl=250 time=0.902 ms
64 bytes from 192.109.197.226: icmp_seq=1 ttl=250 time=0.885 ms
But then, I didn't believe the claim anyway, and I don't need speed. At least I have connectivity. Hopefully the link will stay up.
Notworking: found the fault
|
Topic: Stones Road house, technology, opinion | Link here |
After installing the powerline adapters, back to IRC. Juha Kupiainen had a comment about this photo:
<juha> groggy, this looks wrong
https://lemis.nyc3.digitaloceanspaces.com/grog/Photos/20190923/small/RJ45-4.jpeg
OK, what's wrong with it?
|
I had mainly been looking for mechanical damage, but on closer examination I could only see 4 out of 8 cables. Surely it couldn't have worked like that? No, a problem with the photo; I should have taken it against a dark background. And the sequence matches TIA/EIA-568 T568A:
But that's not what Juha was talking about: he saw past that and saw these markings on the second set of wires:
|
The stripes are longitudinal, but they're clearly visible: the wires are there after all. And wire 3 has a blue stripe, and wire 5 has an orange stripe. WRONG! They should have been the other way round. Over 4 years and I keep discovering blunders made by Jim Lannen! As I wrote:
* groggy sticks another pin into the effigy of Jim Lannen.
But is that the cause? It worked for over two years. Could I have fixed the problem by replacing the connector? First I need to understand where the other end is, and that's still a mystery. And if the other end is also wrong (very likely), then it won't help much. For the time being I'm just happy that I have connectivity again.
Wednesday, 25 September 2019 | Dereel | Images for 25 September 2019 |
Top of page | ||
previous day | ||
next day | ||
last day |
More wildflowers
|
Topic: gardening, opinion | Link here |
Walking the dogs, saw these flowers down Grassy Gully Road:
|
I've seen them before, but somehow I've confused them with Oxalis. They're not that, and it's interesting that there's also a white variant:
|
Time to take a macro lens with me.
ND filter arrives
|
Topic: photography, opinion | Link here |
My 10x neutral density filter, which I needed for the Analemma photos, has arrived. Interestingly, it lets almost exactly the same amount through as the infrared filter. Both of these were taken with P (automatic exposure) mode, and in each case they ended up 1 s at f/4 (EV 4.0, and that handheld!):
|
|
By contrast, without a filter it was 1/250 s at f/8 (EV 14). So at least I have confirmed the accuracy of the 10 EV reduction. But look at the corners! That's really severe vignetting!
Should I send the filter back? This is an ultra wide angle lens, and it's clear that the light in the corners goes through the filter at an angle, so there's more glass to go through. Until proof of the contrary, this could just be a limitation of severe ND filters. And DxO PhotoLab can fix it:
|
Well, up to a point. Increase the correction and the bottom corners get too light, but for my taste the bottom is still too dark. Still, fine for my purpose.
New display card
|
Topic: technology, multimedia, opinion | Link here |
Finally the new display card for teevee has arrived, only 3 weeks since I first ordered one. It's a no-name Nvidia GT730 card, a detail that the packaging tries to compensate for by spraying all kinds of logos, including “Windows” 10 and various features I've neither heard of nor want to know.
The old card was a GT710 from Gigabyte. But they're all the same, right? Yes, plug it in and it worked. But it's noisy! It seems that the fan runs all the time, while on the Gigabyte it only ran when it got hot, which appears not to be very often.
Visit from Lorraine Carranza
|
Topic: general, gardening, opinion | Link here |
Lorraine Carranza along in the afternoon for a brief visit, bringing a Protea with her, a cultivar of an unspecified species (which proved to be Leucadendron salignum) called “Devils Blush”. That's very kind of her, but I wish she wouldn't: we're not really gift people. Still, I managed to find a couple of cuttings of this unidentified bush and pot them for her:
|
|
That's the one that has grown beyond recognition since planting (but not documenting) in April 2018, She's going to try to find somebody to identify them. Here's what the bed looked like then:
|
Ring flash attachment
|
Topic: photography | Link here |
Ten years ago I bought a ring flash adapter for my camera:
|
Recently somebody was asking about it. But the photos I took at the time were really not good, so I tried again:
|
|
It's surprising that these things seem to have disappeared from the market.
Thursday, 26 September 2019 | Dereel | Images for 26 September 2019 |
Top of page | ||
previous day | ||
next day | ||
last day |
Bloody NBN!
|
Topic: technology, general, opinion | Link here |
Into the office this morning to find that we had been off the net since 07:09:10. National Broadband Network scheduled outage? Of course. They have just about the whole working week, every week, as a “scheduled outage”.
But then it occurred to me: I had seen people working on the tower in Enfield when I went past on Tuesday. They even had a backhoe. Were they installing fibre? Dream on! And then Yvonne saw them working both in Enfield and in Dereel yesterday. So out to take a look:
|
|
|
Not exactly a hive of activity, but lots of cars, so maybe they were just out of sight.
Checking on the saved outage messages, it seems that this week was a week of 8 hour outages. And today they pretty much stuck to that: the net came back at 14:31:44, after 7 hours, 22 minutes.
Will they ever stop?
Analemma again
|
Topic: photography, opinion | Link here |
So now I have the right filter for my camera to take the Analemma photos, and the sun was shining. Only I missed the time, so I didn't get a comparison photo to this one:
|
|
But why the oval image? Camera? Lens? Image position? This was with the E-30 and the Zuiko Digital ED 9-18 mm f/4.0-5.6. How does it work with other combinations? Tried it first with the same lens on the E-M1, unfortunately not noticing that the sensitivity was set to 800/30° ISO. But the results were more circular:
|
|
Maybe because it was centred vertically?
Then with the E-30 and the Zuiko Digital ED 12-60 mm f/2.8-4.0 SWD, both at 12 mm and 60 mm. Arguably there's some distortion at 12 mm, but not at 60 mm (and thus not shown):
|
|
So what's the issue? The sensor? The position in the image? I'll have to do some more testing, probably indoors with more controlled circumstances.
Nikolai runs free
|
Topic: animals, opinion | Link here |
After the fiasco with the Everetts 4½ years ago we've kept our dogs on leads at all time when walking them. But it it really necessary? We're not so concerned about people like the Everetts as about kangaroos. On the other hand, the dogs like to investigate things, and they're no longer as active as they were when they were puppies.
Yesterday I walked them by myself, since Yvonne was shopping. She returned when I was on the way back, about 200 m from the house. Nikolai wanted to follow her. OK, off the lead and observe him. He trotted happily away and returned home.
OK, how about a bit more today? Up towards Westons Road, where we seldom see kangaroos, and left him off the lead for most of the way. Things went well, but towards the end we thought it prudent to put him back on the lead because of the dogs up there. On the way back, took him off the lead at the same place. Off he went under the fence into the nearest paddock.
Problem: electric fence. Howl! And we needed quite a bit of effort to get him back again, including myself climbing under the fence. He stuck close by my side for the rest of the way back, so maybe it was a good thing after all.
Another power failure
|
Topic: general | Link here |
Another power failure at 15:30:52, probably only 1 second long.
One wildflower? Many!
|
Topic: gardening, opinion | Link here |
Yesterday I saw some pink wildflowers, which grow everywhere, including on our property. But there's more than one kind. These all grow next to each other:
|
|
|
And of course I forgot my macro lens again.
Friday, 27 September 2019 | Dereel | Images for 27 September 2019 |
Top of page | ||
previous day | ||
next day | ||
last day |
Insurance again
|
Topic: general, opinion | Link here |
A reminder from Non Fleet Recoveries today. Or was it Resilium? No, it was “Fleet Recoveries”. What's the difference?
She reminded me of her previous message, which I had been planning to answer. But looking at it, she had managed to ignore everything I said. Spent quite some time writing a reply, in the process coming across an interesting detail. I had already established that the rear of the car I touched had three separate scratches, only one of which I could have caused:
|
|
The second photo shows the right of the car, the same scratch that's (barely) visible in the first. Three points have become clear:
Put those three details together and there's no way I could have caused any of the damage. Wrote and said so. Since they have no evidence beyond my statement, I'd be interested to see what they can do beyond posturing. I think they know that. One point in particular seems badly thought out:
Regardless, the area that has been impacted has been the only area that has been repaired. Given that the rear bumper bar is one consistent panel, that entire panel needed to be replaced due to the areas of damage that has been caused by your vehicle.
Regardless? Of what? She didn't say. And there's a clear contradiction in that statement. But she did suggest that the repair included more damage than I had caused, because it was not possible not to do so.
Abusive daemons
|
Topic: technology, opinion | Link here |
More spam today. Here from my index:
149 N + 26-09-2019 A??onym0us H??cker To mailer-daemo ( 100) N + IMPORTANT! You h??ve been recorded ??asturbating! I h??ve Mailer Daemon.mp4!
189 N + 26-09-2019 A??onym0us H??cker To abuse@lemis. ( 100) N + IMPORTANT! You h??ve been recorded ??asturbating! I h??ve Abuse.mp4!
It's stupidity like that that gives hackers a bad name.
Horse teeth and new horse
|
Topic: animals | Link here |
Nick Bompas, the equine dentist, along early this morning, removing a milk tooth from Valeta:
|
|
|
And later CJ Ellis came along, joining the other horses in the paddock while he fixed an electric fence connection:
|
|
More garden work
|
Topic: gardening | Link here |
Mick Solly, the gardener, along again today and removed many weeds and a considerable proportion of the native ground cover that I had asked him to keep. Also planted a number of plants: two different kinds of succulents, the lilac and the coriander that Diane from next door gave us, and the Leucadendron salignum from Lorraine Carranza:
|
|
|
|
Yes, there's another plant right next to the Leucadendron, a Salvia microphylla that will need transplanting once it is vigorous enough.
Saturday, 28 September 2019 | Dereel | Images for 28 September 2019 |
Top of page | ||
previous day | ||
next day | ||
last day |
Huevos a la tigre again
|
Topic: food and drink | Link here |
I had a number of leftovers today that looked like the start of a huevos a la tigre. OK, fake it. For me alone (one portion):
quantity | ingredient | step | ||
170 g | potato, cooked | 1 | ||
30 g | onion | 1 | ||
5 g | garlic | 1 | ||
30 g | cooked ham | 1 | ||
25 g | chorizo | 1 | ||
3 g | fresh red chili | 1 | ||
5 g | salt | 1 | ||
40 g | peas | 3 | ||
2 | eggs | 4 |
Cut the potato into cubes, onion, chorizo and ham into small pieces. Peel garlic.
Fry onion in oil, crush garlic into the mixture.
When the onion is glassy, add successively chili, peas, chorizo, ham and finally potatoes. Mix, salt.
|
Place the mixture in a serving dish and fry eggs in the pan.
|
Put eggs on top of mixture, sprinkle with chili flakes and serve.
|
It didn't taste bad, though it's significantly different from the recipe I had been using. I think the best way to look at this dish is for using up leftovers.
More mystery plants
|
Topic: gardening, photography | Link here |
Spring continues, and today I saw something I have never noticed before:
|
|
I had seen these dark brown fronds, which I had thought were dead grass or similar, but it seems that they're what's flowering. And, of course, I had forgotten my macro lens again.
Sunday, 29 September 2019 | Dereel | Images for 29 September 2019 |
Top of page | ||
previous day | ||
next day | ||
last day |
Happy birthday?
|
Topic: technology, general, opinion | Link here |
Yesterday was my birthday. At 47, there's not much to celebrate, and I received only a single birthday greeting in my email. But today Facebook made up for it, lots of people putting messages on my timeline, which I suppose I shall have to look at some time. It was also nice enough to send me an email so that I would find out.
But why today? The rest of the world is a day behind. For Facebook (UTC-7/UTC-8) it's always yesterday, or at least 17 to 19 hours behind.
No Halls Gap
|
Topic: gardening, opinion | Link here |
Today was the annual Grampians Wildflower Show in Halls Gap, which we visited two years ago. I had been planning to go again, but in the end I decided against it. Last time was rather frustrating—one of the reasons we didn't go last year—and there's no reason to believe that we wouldn't run into issues again. Instead we'll revive the visit to Anakie, in the Brisbane Ranges, where we went 10 years ago. They seem to have shut down the annual exhibition there, but we know where we can find the wildflowers. But that seems to be closer to the middle of October, so nothing at all today.
Well, we have wildflowers here too, just not as many. By chance came across this solitary Dianella revoluta in Westons Road:
|
And on the other side of the road a flowering Drosera, a species that I don't recognize:
|
|
Hand-held focus stacking
|
Topic: photography, gardening, opinion | Link here |
I've been puzzling for some time about the Grevillea rosmarinifolia that grow along Stones Road. Are they the same (sub)species or not? Part of the answer could lie in good photos. But the flowers are quite deep, and getting a good, sharp photo is a challenge.
That's what focus stacking is for, of course, but that requires a tripod. Or does it? I had already tried it hand-held with the Dianella revoluta, with catastrophic results (I later realized that I the camera was set for 2 s flash recharge between each image). This time I tried a 20 shot sequence for processing out of the camera, and the results came out better:
|
|
But the final one had too much motion due to the wind, and it failed badly. Still: are those the same species or not? I tend to think not.
Web site: just do it!
|
Topic: technology, opinion | Link here |
Other things have got in the way of my web site migration, almost as I had feared. But the old site is really glacially slow, to the point where people were getting timeouts. Time to hurry things through.
In fact, with just a little corner-cutting, it only took a little over an hour, less time than it took to write it up.
First, of course, I needed to get mail running. Or at least I thought I did, though it occurred to me that there was more work there than I really needed. In the course of the installation, I installed three additional packages: mutt, qpopper, squid, none of which had anything to do with the web server. Also tried to install bip, which was an issue: the port is currently broken. qpopper required changing /etc/inetd.conf, which had multiple differences. Clearly the thing to do here is to look at the diffs on the old machine. The same applied to /etc/rc.conf.
Back (finally!) to the web configuration. I really need to understand where to put the individual web site configurations. Discovered that I had www.lemis.com both in the main /usr/local/etc/apache24/httpd.conf and also in /usr/local/etc/apache24/Includes/www.lemis.com.conf, and it seems that the latter settings override (probably because they get read in last). That's a loose end I need to tie up.
That all seemed to work relatively well. I needed to create the log directory /var/log/www, and I'll have to get Chris Bahlo's stuff in when she has time to look at it.
I did get one strange error message:
[Sun Sep 29 05:23:42.754946 2019] [php7:notice] [pid 43076] [client 167.179.139.35:17989] PHP Notice: Undefined variable: php_errormsg in /home/grog/www.lemis.com/grog/php/includes/onephoto.php on line 1370, referer: http://w3.lemis.com/grog/diary.php
A bit of investigation showed that $php_errormsg is “deprecated”, and should be replaced by appropriate calls to error_get_last (). Interestingly, this was the only place where I had used it, and it only occurred because I hadn't completely synced the site from the old web site.
But apart from that, all was well. OK, bring the site up to date with a sync:
=== grog@www (/dev/pts/2) ~ 5 -> rsync -lKzzavP --exclude=big --exclude=small --exclude=tiny www.lemis.com w3:
One more issue, with SaladoPlayer, an old, worn-out pano program. It still seems to work, but it probably didn't on the old www because some symlinks were broken. To get it to work on either system, I need at least:
=== grog@www (/dev/pts/2) ~ 6 -> ln -s ~/www.lemis.com/grog ~/public_html
But that appears not to be enough. I'll need to check further.
FINALLY: Cut over DNS. Update the A record for www to point to the new site, and add an A record oldwww pointing to the old site. We're done!
Well. No more mail. My fetchmail config talked to www, but mail is still running on the old system. And similar considerations apply to my IRC proxy. Add RRs for mail.lemis.com and irc.lemis.com pointing to the old site, and modify the scripts that connect to the site. Now it really seems to be done.
In passing, there are some strangenesses about DNS. First, people have really removed any trace of my favourite query program, nslookup. There's not even any port available. And restarting named does strange things:
=== root@www (/dev/pts/0) /usr/local/etc/namedb 31 -> service named restart
named not running? (check /var/run/named/pid).
named already running? (pid=806).=== root@www (/dev/pts/0) /usr/local/etc/namedb 32 -> ps aux|grep named
bind 806 0.0 3.0 111568 14844 - Ss 11Aug18 461:39.52 /usr/local/sbin/named -u bind -c
root 59916 0.0 0.3 14644 1268 0 DL+ 6:24AM 0:00.00 grep named
I've seen that before, but that is on eureka, which isn't really correctly configured. This is on a brand-new installation. Killing named works as expected, and then service can restart it, but not stop it again.
And then traceroute
=== grog@eureka (/dev/pts/24) ~ 28 -> traceroute www
traceroute to www.lemis.com (66.42.97.229), 64 hops max, 40 byte packets
1 radiation-tower.aussiebb.net (167.179.136.1) 58.823 ms 24.988 ms 40.366 ms
2 w3 (66.42.97.229) 285.466 ms 200.226 ms 184.692 ms
w3 is at least 10 hops away. What happened there?
Back to the server. First thing, of course, was to look at the load on the systems:
oldwww: last pid: 60006; load averages: 15.08, 17.72, 20.93 up 413+18:53:54 06:34:20
119 processes: 13 running, 106 sleeping
CPU: 90.1% user, 0.0% nice, 9.5% system, 0.4% interrupt, 0.0% idle
www: last pid: 43668; load averages: 0.32, 0.56, 0.56 up 34+00:49:17
59 processes: 1 running, 58 sleeping
CPU: 2.3% user, 0.0% nice, 0.8% system, 0.0% interrupt, 96.9% idle
OK, it takes a while for the DNS entries to refresh—in my case, 30 minutes. But by evening things hadn't changed much.
And now? Still to do:
Alastair Boyanich visits
|
Topic: general, technology | Link here |
Alastair Boyanich (IRC nick uridium) dropped in with his friend Mick, literally for a cup of coffee. Well, Mick drank tea, but they only stayed about 20 minutes. They're on their way from Perth to Sydney via Melbourne. Managed to get rid of the ancient laptops that I found on Monday: Alastair is a collector of old computing hardware, and particularly the Dell prototype (made about 1995) pleased him.
Monday, 30 September 2019 | Dereel | Images for 30 September 2019 |
Top of page | ||
previous day |
Web server progress
|
Topic: technology, opinion | Link here |
In the office this morning, first looked at the web server load.
new: last pid: 43838; load averages: 0.27, 0.39, 0.47 up 34+01:06:03 06:50:37
61 processes: 1 running, 60 sleeping
CPU: 0.0% user, 0.0% nice, 0.8% system, 0.8% interrupt, 98.5% idle
Mem: 53M Active, 1391M Inact, 21M Laundry, 392M Wired, 144M Buf, 126M Free
Swap:
old: last pid: 60303; load averages: 11.16, 14.44, 17.15 up 413+19:10:36 06:51:02
101 processes: 8 running, 93 sleeping
CPU: 94.1% user, 0.0% nice, 5.9% system, 0.0% interrupt, 0.0% idle
Mem: 124M Active, 21M Inact, 322M Wired, 4288K Cache, 58M Buf, 156K Free
Swap: 1024M Total, 201M Used, 823M Free, 19% Inuse, 1864K In, 2588K Out
Why is the old server still so loaded? In the course of time it dawned on me: I don't just have http://www.lemis.com, I also have http://people.lemis.com and http://lemis.com. Gradually updated the DNS for those systems, but things still didn't change very quickly.
Misery Creek Road again
|
Topic: gardening | Link here |
Off to Misery Creek Road in Enfield State Park to walk the dogs today. They had a lot of waiting to do. There wasn't much of interest to be seen on Misery Creek Road itself, much as this time last year, so we parked on the edge of Orchid Track and walked down most of it, while I picked up some photos. There's probably nothing I haven't seen before, though I think this is the first time I've seen these orchids down Orchid Track (a name that I had thought was singularly inappropriate):
|
|
They're clearly some kind of Diuris, but which? In the past I have identified Diuris sulphurea, but these look smaller. I need to check what species are local here.
Apart from that, there was an Acacia species that seems to live on the ground:
|
|
Then more Droseras, this time with clearly pink flowers, though I'm not completely convinced that these two are the same species:
|
|
Then Wurmbea dioica, which I've seen before, but it took me a while to identify them:
|
And of course Epacris impressa. I've seen them in red, white and pink before, but this is the first time I've seen them this close to each other.
|
|
|
And then there are these flowers, which look vaguely like pansies, but the foliage doesn't match:
|
|
Do you have a comment about something I have written? This is a diary, not a “blog”, and there is deliberately no provision for directly adding comments. It's also not a vehicle for third-party content. But I welcome feedback and try to reply to all messages I receive. See the diary overview for more details. If you do send me a message relating to something I have written, please indicate whether you'd prefer me not to mention your name. Otherwise I'll assume that it's OK to do so.
Top of page | Previous month | Greg's home page | Today's diary entry | Next month | Greg's photos | Copyright information |