| cmaloney | Oh that's fun | 00:06 |
|---|---|---|
| Mooncairn | I found a site with S20 specs. Have no idea if they're definitive. | 02:41 |
| Mooncairn | Supposedly, "ECC [is] supported by processor" and it comes with ECC memory. | 02:42 |
| Mooncairn | Haven't been able to track down a BIOS guide yet. | 02:42 |
| Mooncairn | Can confirm ECC mention in user guide. Unfortunately, only a single reference with no info on turning it on or off. | 02:49 |
| Mooncairn | With all cores turned on and almost 6 hours of testing (1 pass complete and 2nd is not quite done), no errors. | 02:50 |
| Mooncairn | Unfortunately, turning all cores on makes the CPU run much hotter. I've got my window open just to keep CPU temp down to 75C. | 02:50 |
| Mooncairn | Whatever's going on, it's sporadic enough that I didn't really notice issues until, what?, a week or two after ditching KDE and switching to GNOME. | 02:52 |
| cmaloney | ye gads | 04:26 |
| Mooncairn | I'm at a loss as to what to do with my computer. | 14:32 |
| jrwren | have you reinstalled OS or tried a different OS? maybe part of libc is corrupted? Have you installed thermal sensors, maybe it is overheating? | 14:35 |
| Mooncairn | Yes, I switched from Debian to Ubuntu. Completely fresh install. | 14:35 |
| Mooncairn | The CPU has thermal sensors. What other kind is there that I could install? | 14:36 |
| jrwren | no idea, but 75C does seem high for idle. what kind of CPU is it? | 14:37 |
| Mooncairn | The CPU can get close to 80C when all 8 virtual cores are running on a task. I try to configure things to not use all 8 cores at once. | 14:37 |
| Mooncairn | The 75C is during a test performed by memtest. | 14:37 |
| Mooncairn | Normally, the CPU idles around 38C. | 14:37 |
| Mooncairn | Some of memtest's tests actively uses only one core and have the rest in a busy loop, which I suspect generates a lot of heat. | 14:39 |
| Mooncairn | Last night I disabled hyperthreading in the BIOS and restarted memtest with only the 4 primary cores. Max temp is 77C during tests. | 14:40 |
| Mooncairn | With the bedroom window closed. ;-) | 14:40 |
| jrwren | I've no idea. Sounds like a bummer to me. | 14:40 |
| Mooncairn | Yes, it is. | 14:41 |
| Mooncairn | I'll have to try booting into Ubuntu's emergency mode and see if some of the errors I was seeing Monday were recorded in the logs. | 14:43 |
| Mooncairn | They all looked memory-related, IIRC. | 14:43 |
| jrwren | maybe memory is bad but memtest just isn't finding it? | 14:48 |
| jrwren | are there multiple dimms? have you tried eliminating one of them? | 14:49 |
| Mooncairn | There's 6 dimms. I haven't tried removing any of them yet. | 14:49 |
| Mooncairn | Nothing in kern.log. In syslog, a few failed assertions in gnome-terminal close to the time of the problems. | 14:51 |
| Mooncairn | I can see Ubuntu gearing up for a crash report on gnome-terminal, which checks out with what I remember happening. | 14:52 |
| Mooncairn | The thing in the log immediately before the crash notification is systemd starting the chezmoi snap. | 14:52 |
| Mooncairn | Nothing about the other GNOME errors that were popping up, and no ~/.xsession-errors file to check. | 14:57 |
| Mooncairn | So... Pull a DIMM (or two? not sure if these are paired), wait up to two weeks to see if anything else happens? | 14:58 |
| Mooncairn | Ok, this is a wild shot in the dark, but I've got nothing better: Could a font cause gnome-terminal and gnome-shell to go nuts? I had very recently installed some fonts in my .local/fonts dir around the time things started to go wild. | 15:20 |
| Mooncairn | So, (malicious?) software problem after all? (Yea, I know it's a bit of a reach.) | 15:20 |
| jrwren | yes, it is possible, but you said it was messed up even with previous distro. | 15:24 |
| Mooncairn | There are LOTS of mentions to the fonts in the syslog. Failure to create cairo scaled font and out of memory errors for gnome-terminal. | 15:26 |
| cmaloney | The OOM issues are a bit worrying | 15:26 |
| Mooncairn | I never was able to pin down what was wrong with the previous distro. Just lots of small random things breaking. | 15:26 |
| Mooncairn | "scaled_font status is: out of memory" | 15:27 |
| Mooncairn | A bunch of that in the log | 15:27 |
| cmaloney | Do you have a spare drive to reinstall onto? | 15:27 |
| Mooncairn | Not really. | 15:28 |
| cmaloney | try this: create a new user | 15:28 |
| cmaloney | and start with that | 15:28 |
| cmaloney | copy things over from your old home dir as needed | 15:28 |
| cmaloney | see if that fixes things | 15:28 |
| cmaloney | barring that, move your home dir to old_home and start fresh | 15:29 |
| Mooncairn | That's what I've been doing with the new install. | 15:30 |
| Mooncairn | I haven't really moved anything from my old home yet. | 15:30 |
| Mooncairn | I was slowly building up my new home from scratch. I intended to keep as much config as possible in git with scripts to install the packages I use. | 15:31 |
| Mooncairn | Okay, I'm fully booted up and logged in. I did rename the fonts files with a .bak extension. One of the ttf files was invalid anyway. It had the diagnostic output from either curl or wget (I forgot which I used) rather than the actual font data. | 15:35 |
| Mooncairn | First thing off the bat is I get another system warning/crash. | 15:35 |
| Mooncairn | "aptitude-curses has stopped unexpectedly" | 15:36 |
| Mooncairn | I hope that's just left over from Monday. I did see that error about aptitude-curses in the syslog. | 15:36 |
| Mooncairn | I'm going to assume that's a left over for the moment. | 15:38 |
| Mooncairn | I'm thinking about sliding the new font(s) back into place and see if everything goes to hell again. | 15:39 |
| Mooncairn | Ok, that was wishful thinking. Ubuntu is reporting errors with gnome-shell. | 15:40 |
| Mooncairn | SIGSEGV, strangely my desktop is still functioning (so far). | 15:44 |
| Mooncairn | syslog and timestamp on the bogus ttf file shows that the creation of the file and things going to hell started within the same 60 second period. | 15:50 |
| Mooncairn | So, correlation but not necessarily causation? The problems noted by gnome-shell started about two minutes afterward. I had had enough time at that point to search duckduckgo for information on signal 7 (BUS). | 15:53 |
| Mooncairn | Perhaps as much as 6 minutes, actually. | 15:53 |
| Mooncairn | The "out of memory" errors are exclusively related to the fonts. I can find no other OOM messages in Ubuntu's logs or in my old Debian install's logs. | 15:56 |
| Mooncairn | Ok, it's definitely the fonts. | 16:01 |
| Mooncairn | I just slid one of the TTF files back into place, changed gnome-terminal's font to it, and BOOM... | 16:02 |
| Mooncairn | The gnome-shell errors on Monday still don't necessarily make sense, though. The ttf file that gnome-terminal blew up on today had been installed a full 13 minutes before things went sideways. | 16:07 |
| Mooncairn | Okay, I guess all I can do right now is to keep monitoring things w/o the offending font files and hope the fonts were the problem. (I'm not sure if I fully believe that, but it's all I have right now w/o further evidence.) | 16:10 |
| Mooncairn | Thanks for bearing with me on this. | 16:11 |
| cmaloney | No problem. | 16:16 |
| cmaloney | Hoping that fixes things | 16:17 |
| cmaloney | OOM and corrupted files are no fun to diagnose | 16:17 |
| Mooncairn | Okay, | 18:45 |
| Mooncairn | Reinstalled the fonts and installed the remaining, this time downloading through Firefox and then installing through GNOME's font manager rather than manually downloading to ~/.local/share/fonts. | 18:46 |
| Mooncairn | So far, so good. gnome-terminal has not crashed. | 18:47 |
| cmaloney | nice | 19:22 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!