| Summary: | dhcpd dies after a few hours with SIGPIPE, systemd-journal segfaulted. | ||
|---|---|---|---|
| Product: | Mageia | Reporter: | Richard Neill <mageia> |
| Component: | RPM Packages | Assignee: | Mageia Bug Squad <bugsquad> |
| Status: | RESOLVED FIXED | QA Contact: | |
| Severity: | critical | ||
| Priority: | Normal | CC: | dmorganec, guillomovitch, lists.jjorge, mageia, thierry.vignaud |
| Version: | Cauldron | Keywords: | NEEDINFO |
| Target Milestone: | --- | ||
| Hardware: | x86_64 | ||
| OS: | Linux | ||
| Whiteboard: | |||
| Source RPM: | systemd | CVE: | |
| Status comment: | |||
| Bug Depends on: | |||
| Bug Blocks: | 2120 | ||
| Attachments: |
Backtrace from core dump
Backtrace |
||
Hi, thanks for reporting this bug. As there is no maintainer for this package I added the committers in CC. (Please set the status to 'assigned' if you are working on it) CC:
(none) =>
dmorganec, guillomovitch, lists.jjorge, thierry.vignaud Please: 1) install gdb systemd-debug (you'll need to enable the Core Release Debug media), 2) run "gdb systemd-journal /path/to/your/core/file" 3) type "bt" 4) copy the backtrace printed by GDB in some text file you'll attach (not paste) to this bug report Keywords:
(none) =>
NEEDINFO Created attachment 1713 [details]
Backtrace from core dump
As requested. Let me know if there is anything else I can do.
You need to install systemd-debug (and glibc-debug and maybe dbus-debug) else the trace is useless
Thierry Vignaud
2012-03-10 18:05:27 CET
Attachment 1713 is obsolete:
0 =>
1 I did install systemd-debug already. I've just installed glibc-debug and debus-debug, and repeated your instructions. I still get the same output. What should I do? Created attachment 1725 [details]
Backtrace
I realise why gdb had trouble: systemd-journald isn't in $PATH. Trying the following is more helpful:
gdb /lib/systemd/systemd-journald core.systemd-journald
But I get this error:
the debug information found in "/usr/lib/debug//lib/systemd/systemd-journald.debug" does not match "/lib/systemd/systemd-journald" (CRC mismatch).
which suggests that in the 8 hours between my experiencing the crash and downloading the debug packages, the version of systemd-debug has changed.
I currently have:
systemd-debug-43-4.mga2 and systemd-40-3.mga2 installed.
So, is it possible to fetch the older debug package? Or alternatively, can we assume that the newer version of systemd contains a fix for this isssue?
Is this still a problem with systemd-44? CC:
(none) =>
mageia Ping? Seems to be OK now. It's run for a while without trouble :-) I'll make this as resolved fixed. Thanks for your help. Status:
NEW =>
RESOLVED |
Description of problem: DHCPD dies after a few hours. There is nothing in the logs to suggest why, but it appears to have received a sigpipe. It looks like the cause of this is that journald dumped core, thereby making dhcpd die. I've just installed Mageia beta 2 on a new PC yesterday, updated everything (so running current Cauldron as of a few hours ago), and done some basic config with MCC to set up network connection sharing. It was working for a few hours, then dhcpd died. "service dhcpd status" shows the following: ------ dhcpd.service - DHCPv4 Server Daemon Loaded: loaded (/lib/systemd/system/dhcpd.service; enabled) Active: inactive (dead) since Sat, 10 Mar 2012 04:35:21 +0000; 1h 8min ago Main PID: 22837 (code=killed, signal=PIPE) CGroup: name=systemd:/system/dhcpd.service ------ In dmesg, I see: ----------------- systemd-journal[335]: segfault at 7f3d322655e8 ip 000000000040d99c sp 00007fffec036cd0 error 4 in systemd-journald[400000+1e000] systemd-coredump[21878]: Detected coredump of the journal daemon itself, diverting coredump to /var/lib/systemd/coredump/. ata1.00: configured for UDMA/133 ata1: EH complete ata2.00: configured for UDMA/133 ata2: EH complete EXT4-fs (sda1): re-mounted. Opts: acl,commit=0 EXT4-fs (sdb1): re-mounted. Opts: acl,commit=0 EXT4-fs (sda6): re-mounted. Opts: acl,commit=0 systemd-journal[21880]: segfault at 7fc343292448 ip 000000000040d99c sp 00007fff4676ed80 error 4 in systemd-journald[400000+1e000] systemd-coredump[23321]: Detected coredump of the journal daemon itself, diverting coredump to /var/lib/systemd/coredump/. ------------------ Dhcpd can be restarted successfully. I have a coredump file, 2.1MB. If anyone wants it I can upload it. I've saved a copy of /var/log/*, so ask if you want more info.