Starting with systemd version 189 systemd-journald segfaulted quite often (see an example below). At the moment I don't have systemd with no debugging symbols, so I can't provide a useful backtrace. However, this https://gist.github.com/3659213 might be the same issue. Sep 13 09:05:48 xxxxx systemd-journal[13445]: Journal started Sep 13 09:05:48 xxxxx systemd[1]: systemd-journald.service: main process exited, code=dumped, status=11 Sep 13 09:05:48 xxxxx kernel: systemd-journal[153]: segfault at 7fa450257480 ip 0000000000416ebb sp 00007fff95521b10 error 4 in systemd-journald[400000+28000] Sep 13 09:05:48 xxxxx systemd-coredump[13443]: Detected coredump of the journal daemon itself, diverting coredump to /var/lib/systemd/coredump/. Sep 13 09:05:48 xxxxx systemd[1]: Unit systemd-journald.service entered failed state. Sep 13 09:05:48 xxxxx systemd[1]: systemd-journald.service holdoff time over, scheduling restart. Most likely not useful output, but anyway: Core was generated by `/usr/lib/systemd/systemd-journald'. Program terminated with signal 11, Segmentation fault. #0 0x0000000000416ebb in ?? () (gdb) where #0 0x0000000000416ebb in ?? () #1 0x000000000041710b in ?? () #2 0x00000000004185e9 in ?? () #3 0x0000000000406d8d in ?? () #4 0x0000000000409f2b in ?? () #5 0x000000000040a2a4 in ?? () #6 0x000000000040a3d1 in ?? () #7 0x000000000040463c in ?? () #8 0x000000368922460d in __libc_start_main () from /lib64/libc.so.6 #9 0x0000000000404fed in ?? () #10 0x00007fff955226c8 in ?? () #11 0x00000000ffffffff in ?? () #12 0x0000000000000001 in ?? () #13 0x00007fff95522ea8 in ?? () #14 0x0000000000000000 in ?? ()
Hmm, any chance you can install debug symbols and regenerate the backtrace?
I'm using Gentoo, so I have to recompile systemd with debug symbols. I suppose I have to have coredump from the systemd-journald with debugging symbols, right. If so, reboot is needed, but I can do so only tomorrow (now I'm quite busy). If reboot is not needed, I can do this quite quickly. Pls advise.
I made some changes to git now. Would be cool if you could test this whether this makes the probs go away. Thanks! If it doesn't I'd appreciate a backtrace of the failures.
(In reply to comment #2) > I'm using Gentoo, so I have to recompile systemd with debug symbols. I suppose > I have to have coredump from the systemd-journald with debugging symbols, > right. If so, reboot is needed, but I can do so only tomorrow (now I'm quite > busy). If reboot is not needed, I can do this quite quickly. Pls advise. If the changes from git don't help, then yes, please compile systemd with debug symbols, make it crash and generate a backtrace from the coredump.
(In reply to comment #4) > If the changes from git don't help, then yes, please compile systemd with debug > symbols, make it crash and generate a backtrace from the coredump. Lennart, Could you point me to this particular commit?
I am quite sure this is fixed with 191. Please check if you can reproduce this with 191, if so open a new bug with a bt from 191. Thanks!
(In reply to comment #6) > I am quite sure this is fixed with 191. Please check if you can reproduce > this with 191, if so open a new bug with a bt from 191. Thanks! Thanks! I've installed systemd-191 only this morning, no crashes so far. Hopefully everything will be fine.
Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.