#1

People,

I posted this note on the general Fedora list a little while ago - I got one response but couldn’t make any more progress debugging F29/30 x86_64 XFCE WS):

"I have been having fairly regular hangs of the whole system - sometimes I have a little warning of a rapidly slowing system and I can systematically exit from all the apps and reboot cleanly - but mostly not and I have to hit the reset button. I used to think it was overloading the (fast) system (with 32GB RAM) with too many Chrome windows and tabs but it happens occasionally when the system is very lightly loaded too. I have new RAM and have checked it with the Pro version of MemTest86 a couple of times so I don’t think it is RAM. I have pretty much replaced all the other bits of hardware now except the MB.

Any suggestions about further debugging would be greatly appreciated!"

I have been interested in Atomic / SB for a while now and have been using Atomic on a big disk backup server for somwtime but now I am thinking that running a WS SB system and containerising all apps - which might help me isolate where the hanging problem is (assuming it is in software . . )? Using containers might cause only a single container to hang instead of the whole machine?

What do people think?

Thanks,
Phil.

#2

Not quite, containers will still share your core resources so I don’t think they’d help much here. You can limit them, but given we don’t know what the problem is there’s not much to limit.

Personally, I use this GNOME extension to keep track of my system resources. I’d advise to try something similar, or even just leave System Monitor running in the background but take careful note of CPU/RAM usage and most notably disk I/O usage (which can often be ignored but can cause problems).

Also, after a force reset, try journalctl -b-1 -e to see the last things that were logged before the poweroff.

#3

@refi64 ,

Thanks for the response!

Right.

CPU/RAM was always OK but I will have a closer look at disk I/O.

I have been collecting the tails of /var/log/messages up until crashes but they don’t indicated anything unusual unfortunately . . I will try the journalctl command but that should give much the same info?

I have now been playing with SB on my laptop and tonight on my workstation and as I might make progress, will keep watching for the old hanging problem however ATM I have another Q about keeping my old /home stuff - but I will put that in a new post.

#4

journalctl pretty much supercedes most traditional text logs and generally contains far more info than any leftover text logs might say.

#5

Ah OK - cool, I will give that a go - thanks!

#6

@refi64 ,

Looks like you were right:

I left SB running last night - when I got back to it this morning, it had hung too . . I hope I am not stretching the friendship too much but I attach the (short) output from your command - if you have time to look at it? (you were more helpful than the single response I got from the Fedora mailing list!) - maybe chrome is a problem?

Thanks,
Phil.

#7

Sadly those logs don’t seem to have much interesting, a few perf messages seem to imply something is eating CPU a bit but it didn’t look like anything severe.

I should’ve probably asked this to start:

  • If this is a prebuilt system, what model is it?
  • Would you happen to have an nvidia GPU? If so, are you on the proprietary driver?
  • Have you ever tried any other distros to know if this occurs there?
#8

@refi64 ,

Damn . . that is what I thought . .

Yep - I have been building my own machines since 1990 . . I have used memtest86 on the RAM which seems OK and have swapped out the PSU with no improvement so it might be a flakey MB problem . .

Integrated Graphics Processor- Intel® HD Graphics support

I haven’t used another distro as a day machine since RH4 (ie pre-Fedora).

#9

Hmm…

I’m guessing at this point that this is some sort of a driver issue, since nothing interesting was happening anywhere else. Does journalctl -b-1 -e -k show anything? (It’s basically the same command as before, but with -k at the end.)

#10

@refi64 ,

Well I have reinstalled from scratch since the hang (see my posts in the XFCE thread) but I will try the “-k” next time and come back here.

Thanks for the help - I appreciate it!

Phil.

1 Like