systemdlete | firefox and thunderbird are crashing after I rebooted all my systems, on both VMs and hosts. This behavior is new, only since the reboot to new kernel (5.10.0-16). Anyone else having this, or knows what is going on? | 21:09 |
---|---|---|
systemdlete | was using 5.10.0-14 previously (yes, I never ran 5.10.0-15, sorry) | 21:10 |
systemdlete | Googling, I found a hit for the error messages from 2018 and earlier, but nothing more recent. | 21:11 |
systemdlete | the solution then (at least, one solution anyways) was to revert to an earlier version of ff. | 21:12 |
fsmithred | was that after a firefox update or a kernel update? | 21:13 |
systemdlete | Is it time, once again, to switch to chromium | 21:13 |
fsmithred | ewwww | 21:13 |
systemdlete | fsmithred, well, not sure, there have been updates of both recently | 21:13 |
fsmithred | I mean the old post | 21:13 |
systemdlete | yes, it was | 21:13 |
rwp | systemdlete, is this in Ceres Unstable? | 21:13 |
systemdlete | no chimaera | 21:14 |
fsmithred | I'm upgrading now. Kernel is in the list, but firefox is not and has been working ok. | 21:14 |
systemdlete | fsmithred, upgrading from _____ to ____ ? | 21:14 |
systemdlete | (just wondering if it's the same sequence as mine) | 21:14 |
fsmithred | upgrading to current chimaera from current chimaera minus 64 package. | 21:15 |
fsmithred | -15 to -16 | 21:15 |
systemdlete | ah, ty | 21:15 |
systemdlete | note that I skipped -15; went directly from -14 to -16 | 21:15 |
fsmithred | rebooting | 21:15 |
rwp | I have not had any problems with Firefox on Chimaera with Linux 5.10 and firefox-esr 91.11 | 21:15 |
systemdlete | it's crazy, because in my case, both tbird AND ff are crashing. | 21:15 |
systemdlete | I had not either, until just a little while ago when I rebooted to the new kernel (and perhaps newer ff and tbird, not 100% sure) | 21:16 |
rwp | I think that is a clue that it is in the libraries that both of those share using. Both being Mozilla and all. | 21:16 |
systemdlete | rwp: right | 21:16 |
systemdlete | I've seen stuff just like this in the past. Iirc, I had to wait for a fix from upstream | 21:17 |
systemdlete | (that was on centos, not devuan, though) | 21:17 |
systemdlete | (years ago) | 21:17 |
systemdlete | rwp, fsmithred: I'm using only stock (repo) versions of the kernels and ff and tbird. Just for further reference/context. | 21:18 |
systemdlete | Also, this is happening on the StarLinux variant also. (but let's not go THERE, ok? LOL) | 21:19 |
systemdlete | I am not certain but Star might be stuck at beowulf. | 21:19 |
systemdlete | I have not seen an update since then. | 21:20 |
fsmithred | you upgraded it to chimaera? | 21:20 |
fsmithred | I opened 12 tabs. What else do I need to do? | 21:20 |
systemdlete | 4.19.0 is the kernel running on StarLinux | 21:21 |
systemdlete | and I've kept it up to date | 21:21 |
fsmithred | up-to-date beowulf or chimaera? | 21:21 |
fsmithred | apt policy libc6 please | 21:21 |
systemdlete | I think, but not sure, StarLinux was beowulf. | 21:21 |
fsmithred | I can tell you if you answer. | 21:22 |
systemdlete | 2.31-13+deb11u3 | 21:22 |
systemdlete | that's on chimaera systems. | 21:22 |
fsmithred | yes | 21:23 |
fsmithred | chimaera | 21:23 |
fsmithred | any particular website makes it crash? | 21:23 |
systemdlete | on Star, it is 2.28-10+deb10u1 | 21:23 |
systemdlete | (and 12 tabs is plenty to repro) | 21:23 |
fsmithred | that's beowulf | 21:23 |
systemdlete | right, as I thought... | 21:23 |
fsmithred | you have the problem on both systems? | 21:24 |
systemdlete | so kernel level might be irrelevant ? | 21:24 |
systemdlete | yes | 21:24 |
fsmithred | yeah, I'm running beowulf most of the time (right here, now) and no browser crashes. | 21:24 |
systemdlete | hmmm | 21:24 |
systemdlete | ok | 21:24 |
systemdlete | so it's just me then | 21:24 |
fsmithred | also t'bird on this beowulf. | 21:25 |
rwp | I have my new machine running Chimaera and no crashes. systemdlete, Try opening firefox from the command line with: firefox -safe-mode | 21:25 |
systemdlete | thx for that info | 21:25 |
fsmithred | how does the crash act? | 21:25 |
systemdlete | rwp: sure... hold on | 21:25 |
systemdlete | fsmithred, I get the crash dialog | 21:25 |
fsmithred | oh | 21:25 |
rwp | Upstream docs on that mode: https://support.mozilla.org/en-US/kb/diagnose-firefox-issues-using-troubleshoot-mode | 21:25 |
systemdlete | it offers to restart it | 21:25 |
systemdlete | safe-mode seems to alleviate it, yeah rwp... but let me try some more things | 21:26 |
rwp | The next step in the Mozilla Firefox trouble shooting decision tree is: https://support.mozilla.org/en-US/kb/troubleshoot-extensions-themes-to-fix-problems | 21:28 |
systemdlete | I've disabled all extensions. | 21:29 |
systemdlete | Except for the ddg one. | 21:29 |
systemdlete | Here is an exmple of the cmd line trace from ff: https://pastebin.com/3amqvYmW | 21:31 |
systemdlete | (if interested) | 21:32 |
systemdlete | rwp: They suggest disabling hw accel -- I will try that | 21:34 |
systemdlete | but it still crashes, even with hw accel disabled (unchecked box) | 21:36 |
rwp | I really have no idea. I can only walk through the upstream decision tree for debugging these types of problems. | 21:37 |
rwp | But a combination kernel, library, and program upgrades and specific hardware is plausible to trip over a problem of hardware acceleration. Maybe. | 21:39 |
systemdlete | This was not an issue until after I rebooted to this new kernel. I can try rebooting to the old kernel. Bbs... | 21:42 |
systemdlete | rolling back to previous kernel does not help. | 21:54 |
systemdlete | I noticed this message just now: Failed to open curl lib from binary, use libcurl.so instead | 21:54 |
systemdlete | That message does not always seem to show up | 21:55 |
systemdlete | Am I supposed to coerce these programs (ff and tbird) to somehow make sure to use the shared library? | 21:56 |
systemdlete | ah! In the kern.log, I see tons of these: [drm:vmw_msg_ioctl [vmwgfx]] *ERROR* Failed to open channel. | 21:59 |
rwp | Shared libraries such as for curl should Just Work. That doesn't mean that isn't where the breakage is though. | 21:59 |
systemdlete | btw, I should tell you that this is in a VM. I tried ff on the host, and I didn't see problems, but I only did some very basic tests. | 21:59 |
systemdlete | so this might be a video driver issue? Maybe? But I haven't done ANYTHING to vbox since looooong before the kernel reboot | 22:00 |
systemdlete | I wonder... if I revert all the way back to the -14 kernel, maybe that would restore things | 22:01 |
rwp | If someone asks if some particular random thing might be a problem the answer is always yes, yes it might be a problem. | 22:01 |
rwp | But guessing like that is a hard way to debug. It's better to work through from the known data. | 22:01 |
systemdlete | rwp: Well, I *was* running the -14 kernel before these crashes started occurring. | 22:02 |
rwp | But you booted back to the -14 kernel and that did not resolve the problem. | 22:03 |
systemdlete | There is, in fact, a vbox issue with video, so that could be the smoking gun. | 22:03 |
systemdlete | rwp: NO! | 22:03 |
rwp | No? | 22:03 |
systemdlete | actually, the -14 kernel is gone now. I only had -15 to work with. That's what I am running | 22:03 |
systemdlete | So maybe the "rollback" is not a fair test | 22:03 |
systemdlete | If I use apt to install the -14 kernel, will that cause -16 to disappear, or something like that? | 22:04 |
systemdlete | (not sure how apt handles this, sorry) | 22:04 |
rwp | Installing a different kernel with different version number will not make other version'd kernels disappear. | 22:05 |
systemdlete | good, thanks | 22:06 |
rwp | Sometimes though they release updated kernels of the same version to be installed on top of the existing kernels. Those do replace each other. | 22:06 |
systemdlete | do I need just the kernel package, or will I need to also install something else (I forget this) | 22:06 |
rwp | I don't approve of that since it can cause problems depending upon how they handle things. If they get it right then fine. But too often they have not. | 22:06 |
rwp | Do you still have the linux-image .deb file in your /var/cache/apt/archives directory? | 22:07 |
systemdlete | so I shouldn't try installing -14? | 22:07 |
rwp | If so then you could install it again and try it again. | 22:07 |
systemdlete | I will look... | 22:07 |
fsmithred | I don't have any old kernel debs. Just other stuff. | 22:07 |
systemdlete | no kernel debs in my archives either (except other pacakges) | 22:09 |
fsmithred | looks like everthing back to -11 is still in the repo | 22:09 |
rwp | I don't happen to have that kernel around either. I also don't see it at snapshot.debian.net making me wonder... | 22:10 |
systemdlete | sadly, I don't have a VM snapshot either... | 22:10 |
systemdlete | (sometimes I do have those laying around) | 22:10 |
fsmithred | I just downloaded -11 | 22:11 |
rwp | Correction, sorry, http://snapshot.debian.org/ doesn't list it either. | 22:11 |
systemdlete | When I tried to install -14, I got a bunch of hash mismatch error messagse | 22:11 |
fsmithred | I can't try to intsall -14 because it's still there. | 22:12 |
systemdlete | this time, maybe it is working... | 22:12 |
systemdlete | flakey, if so | 22:12 |
rwp | linux-image-5.10.0-14-amd64 definitely existed because I had installed it on 2022-05-27. Therefore there must be an archive of it somewhere. | 22:13 |
systemdlete | I wonder if NSO is busy in our repos | 22:13 |
fsmithred | it should work, and it should not remove the -16 kernel. | 22:13 |
systemdlete | again, it does seem to be doing something atm | 22:13 |
systemdlete | oooh. | 22:14 |
systemdlete | I just checked df on my /home and it is under 1G | 22:14 |
fsmithred | that's tight | 22:14 |
systemdlete | I do have about 15 tabs open in the browser on this VM (a chimaera, not star) | 22:14 |
systemdlete | yeah, possibly. | 22:14 |
systemdlete | but I am not seeing errors about space | 22:14 |
systemdlete | just those "channel" errors, which might be indicative of such a thing, idk | 22:15 |
fsmithred | should be more than 5% free | 22:15 |
systemdlete | my /home is about 2.5G, 59% used | 22:16 |
fsmithred | oh | 22:16 |
systemdlete | install of -14 kernel is stuck at 20% for about the last 5 minutes | 22:16 |
systemdlete | (unpacking) | 22:16 |
systemdlete | just advanced, nvm | 22:17 |
fsmithred | took me a minute at least. 100Mb/s down speed here. | 22:17 |
rwp | Ah, found it, you can download that kernel from here: http://deb.devuan.org/merged/pool/DEBIAN-SECURITY/updates/main/l/linux-signed-amd64/linux-image-5.10.0-14-amd64_5.10.113-1_amd64.deb | 22:17 |
rwp | Found from https://pkginfo.devuan.org/cgi-bin/package-query.html?c=package&q=linux-image-5.10.0-14-amd64=5.10.113-1 | 22:17 |
systemdlete | I'm supposedly 400 down | 22:17 |
systemdlete | done! | 22:17 |
systemdlete | ok, give me a moment to reboot. bbs | 22:17 |
rwp | Wait... You already had a deb of it? I missed that. Sorry. | 22:18 |
systemdlete | bak | 22:22 |
systemdlete | testing now under -14 | 22:22 |
systemdlete | crud. crashed | 22:23 |
systemdlete | so not the kernel | 22:23 |
fsmithred | check power supply and memory? | 22:23 |
systemdlete | I have my PCs on UPS's | 22:23 |
fsmithred | I mean check internal voltages and do memtest | 22:24 |
fsmithred | or try ff from a live usb | 22:25 |
systemdlete | VM has 3.84 G memory, using about 1G of it with FF running. Disk space barely budging on /home | 22:26 |
systemdlete | the /tmp has nearly 2G and only using 1% of that | 22:26 |
systemdlete | other fs look ok I think | 22:26 |
systemdlete | it has been very hot here for a few days now and I am not running A/C | 22:29 |
rwp | It's a VM so that seems less likely the problem. I would be inclined to purge --autoremove firefox-esr and perhaps *curl* too since it was implicated then install all again. | 22:29 |
systemdlete | rwp: good idea. Will proceed to do so now. Do you think a reboot is in order? | 22:30 |
systemdlete | I'm thnking: test reinstalled ff and tb before reboot, then try them after rebooting to -16 kernel | 22:30 |
systemdlete | apt purge firefox thunderbird curl gives me: E: The package cache file is corrupted, it has the wrong hash | 22:31 |
rwp | Well... That's a BIG CLUE. | 22:31 |
rwp | I would also look at this list "dpkg -l | grep ^rc" and purge those off to clean up too. (No idea what you will find there.) | 22:31 |
systemdlete | I DID mention that I had gotten a bunch of hash mismatch errors when I first tried to install -14 kernel, but not the 2nd time | 22:32 |
rwp | Maybe you are having RAM problems after all... | 22:33 |
systemdlete | ok... so I could run a memtest but I will of course have to shut this down | 22:33 |
systemdlete | for hours | 22:33 |
fsmithred | probably check smart first. That only takes a few minutes. | 22:34 |
fsmithred | seconds... smartctl -a /dev/sda | 22:34 |
systemdlete | on the host? or vm? | 22:34 |
systemdlete | I mean, the VM does not really support smart | 22:34 |
fsmithred | oh, on the host | 22:35 |
fsmithred | same with the memtest | 22:35 |
fsmithred | for temps, put some temp display on the desktop so you can watch it. But usually, overheating just causes shutdown. | 22:36 |
rwp | Look in the host /var/log/syslog and /var/log/kern.log for anything that looks like problems. | 22:37 |
systemdlete | yeah, just looked through the hosts's kern and message logs | 22:39 |
systemdlete | I know of a few error messages to look for in case of drive issues, like ata error messages. I don't see any of those. | 22:40 |
systemdlete | I don't see any recent messages that indicate hardware issues. | 22:40 |
systemdlete | And I looked back a few hundred lines | 22:40 |
systemdlete | skimmed | 22:40 |
systemdlete | fsmithred, I looked at smart output on host, but I always forget which ones to look for. I usually look for remapped blocks and see if the pool is exhausted | 22:42 |
systemdlete | and thanks to the lack of standards, it is hard to figure which cells are important for a particular drive type | 22:42 |
fsmithred | age and reallocated sectors are good ones to check. | 22:42 |
systemdlete | realloc'd are the sparing ones I mentioned | 22:43 |
systemdlete | (what I meant by remapped) | 22:43 |
fsmithred | yeah, I figured. It should be zero. | 22:43 |
systemdlete | they are, on both drives. But "age" -- many cells are type old_age, but I think that is their type, not their status | 22:45 |
systemdlete | (I'm really not sure) | 22:45 |
fsmithred | power-on hours | 22:46 |
systemdlete | and I would think I'd be seeing those awful console messages indicating terrible problems with the drive(s)--not to mention long lags and knocking noises from the drives, if a problem exists. | 22:46 |
systemdlete | that's been my experience... | 22:46 |
systemdlete | one's around 8,000 hours, and the other about 15,000 hours | 22:47 |
systemdlete | When I disabled the hardware acceleration in FF, I got some really ugly artifacts from the video... and they didn't go away immediately. I'm wondering if my video card might be the issue? | 22:50 |
systemdlete | It's a couple years old, maybe 3. | 22:50 |
_ds_ | That's possible. Are you seeing odd artefacts from time to time? | 22:50 |
_ds_ | (other than those videos) | 22:51 |
systemdlete | _ds_, thanks for helping. Yeah, I notice a lot of "flashes" | 22:51 |
systemdlete | sorry, when I said "video" I meant the screen, not youtube videos etc | 22:51 |
systemdlete | keep in mind, most of my work is inside VMs | 22:52 |
systemdlete | (but not all VMs are using graphical UIs, like appliances) | 22:52 |
_ds_ | Had a problem with one card not long ago – I was seeing some issues with tiling in certain situations. Random blocks of colour overlaid, but a clear pattern – 4×4 tiles, each 8×8 pixels | 22:54 |
_ds_ | That kind of thing Just Happening without changes in relevant software… hardware problem. | 22:55 |
_ds_ | Can also cause odd segfaults… | 22:57 |
systemdlete | so you replaced it? | 22:57 |
_ds_ | Definitely grab another graphics card for testing. | 22:57 |
_ds_ | Yes. | 22:57 |
_ds_ | Problems went away immediately. | 22:57 |
systemdlete | :) | 22:57 |
_ds_ | (At least prices are somewhere nearer sane these days.) | 22:57 |
systemdlete | ok, well when I take this PC down for memtest, I can try installing another card. I think I have a spare here. | 22:58 |
_ds_ | I remember, some years ago, memory suddenly failing. Caused immediate and severe problems with trying to run most things (file systems survived, though); memtest spewed a lot of errors. | 23:01 |
systemdlete | A few months back, I was having some funky problem, and people here suggested memtest. So I did that, but after about 12 hours, no errors. So I rebooted and carried on. A little while later some updates came through and those seemed to fix it all. | 23:02 |
systemdlete | but, hw can be the issue. Speaking of which... | 23:02 |
systemdlete | while I was typing that long trope, I saw my open web page suddenly artifact itself, then refresh. | 23:03 |
systemdlete | again, plenty of memory, and plenty of disk space. | 23:03 |
systemdlete | can't locate the card | 23:04 |
fsmithred | ? | 23:08 |
systemdlete | fsmithred, I meant that I couldn't find that spare video card | 23:09 |
systemdlete | I was certain I had an extra here | 23:09 |
systemdlete | In the old days, I'd hop the bus and proceed directly to Fry's... but they are no more. :( | 23:12 |
systemdlete | I can go to the Best Buy near me... but I hate shopping in that place. That company is a mess. | 23:12 |
systemdlete | I don't even know if they have an ordinary video card that doesn't cost $600 in the store. | 23:13 |
systemdlete | I wonder if moving the card to another pci slot might help. And maybe cleaning the connectors. | 23:14 |
systemdlete | sometimes that also does miracles. | 23:14 |
systemdlete | but only sometimes. | 23:14 |
systemdlete | FOUND IT | 23:19 |
systemdlete | ok | 23:19 |
systemdlete | thanks to everyone who chimed in with suggestions. | 23:19 |
systemdlete | I will shut down the PC now and run memtest while I hike over to Best Buy and see if they have video cards. If not, I will try the one here--the problem is I am not even sure what it is. Probably the only way to know is to insert it and boot with it. | 23:20 |
systemdlete | bbl | 23:21 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!