Bug 149475
| Summary: | ssh-agent started before gnome-session dies | ||
|---|---|---|---|
| Product: | [Fedora] Fedora | Reporter: | Michael Hoffman <hoffman+redhat> |
| Component: | openssh | Assignee: | Tomas Mraz <tmraz> |
| Status: | CLOSED NOTABUG | QA Contact: | Brian Brock <bbrock> |
| Severity: | medium | Docs Contact: | |
| Priority: | medium | ||
| Version: | 3 | CC: | mharris |
| Target Milestone: | --- | ||
| Target Release: | --- | ||
| Hardware: | i386 | ||
| OS: | Linux | ||
| Whiteboard: | |||
| Fixed In Version: | Doc Type: | Bug Fix | |
| Doc Text: | Story Points: | --- | |
| Clone Of: | Environment: | ||
| Last Closed: | 2005-04-06 16:00:13 UTC | Type: | --- |
| Regression: | --- | Mount Type: | --- |
| Documentation: | --- | CRM: | |
| Verified Versions: | Category: | --- | |
| oVirt Team: | --- | RHEL 7.3 requirements from Atomic Host: | |
| Cloudforms Team: | --- | Target Upstream Version: | |
| Embargoed: | |||
| Bug Depends On: | 134494 | ||
| Bug Blocks: | |||
|
Description
Michael Hoffman
2005-02-23 14:38:36 UTC
I cannot reproduce the ssh-agent crash neither when I start X session from gdm nor using startx. Have you changed anything in your X startup scripts? It only happens when I use gdm. I have changed things in my X startup scripts in an attempt to get it to work, but it never worked to start with. Commonly the ssh-agent process will live for minutes or hours, and then die without warning. I never have this problem when starting ssh-agent manually. Is there any way I can provide additional debugging information? Hmm so it crashes only after some time? What ps -efa | grep ssh-agent gives you when it's still running? Also could you post here the exact commandline it's started with? The most interesting thing would be to find out at what exact conditions it dies, also if it dies by SEGFAULT or other means. I started the following script in a full-screen terminal:
while true; do
date
ps -efa | grep [s]sh-agent
sleep 1
done
Then I logged in. The results:
Wed Mar 9 15:09:25 GMT 2005
Wed Mar 9 15:09:26 GMT 2005
Wed Mar 9 15:09:27 GMT 2005
hoffman 5096 1 0 15:09 ? 00:00:00 ssh-agent -s
Wed Mar 9 15:09:28 GMT 2005
hoffman 5096 1 0 15:09 ? 00:00:00 ssh-agent -s
[...]
Wed Mar 9 15:46:08 GMT 2005
hoffman 5096 1 0 15:09 ? 00:00:00 ssh-agent -s
Wed Mar 9 15:46:09 GMT 2005
hoffman 5096 1 0 15:09 ? 00:00:00 ssh-agent -s
Wed Mar 9 15:46:10 GMT 2005
Wed Mar 9 15:46:11 GMT 2005
Wed Mar 9 15:46:12 GMT 2005
So it just died 37 minutes later for no apparent reason. I wasn't doing anything
ssh related at the time.
Because I cannot reproduce it here it would help if you could find out how it dies - if it gets SIGSEGV or simply exits or what exactly happens with it. Also it would help if you could try to reproduce the behaviour with keys which would be generated specially for this purpose -> not security sensitive so you could attach them to the bug report. I've updated xinitrc to 4.0.15 in rawhide. You may want to test with that to see if it has any effect on this problem. Michael, did the xinitrc 4.0.15 from rawhide help? I tried the new xinitrc and it didn't change anything. I also tried OpenSSH
4.0p1, no change. I tried stracing the ssh-agent process and this is what I get
when it dies:
select(4, [3], [], NULL, NULL) = ? ERESTARTNOHAND (To be restarted)
--- SIGTERM (Terminated) @ 0 (0) ---
unlink("/tmp/ssh-UCbLh23175/agent.23175") = 0
rmdir("/tmp/ssh-UCbLh23175") = 0
exit_group(2) = ?
Process 23176 detached
Hmmm nice - so the ssh-agent is getting SIGTERM. Definitely not a bug in ssh-agent that it exits then. Now, you must find who/which evil process is sending the SIGTERM signal to the ssh-agent. If this process is from the openssh package, then reopen this bug, otherwise open a new bug on the package from which it is. |