Red Hat Bugzilla – Bug 67115
xfs crashes when sent SIGUSR1
Last modified: 2007-04-18 12:43:20 EDT
From Bugzilla Helper:
User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:0.9.9) Gecko/20020408
Description of problem:
According to docs, xfs should reread its config file on SIGUSR1 arrival. But
sometimes it crashes. I've spent some time trying to catch the situation when
it happens, and finally found the Anaconda-generated /etc/X11/fs/config on a
just-installed system to give 100% hits (I'll attach the config).
Version-Release number of selected component (if applicable):
Steps to Reproduce:
1.in some window do "strace -p `cat /var/run/xfs.pid`"
2.in another window, do "killall -USR1 xfs"
Actual Results: xfs dies, with strace printing
--- SIGPWR (Power failure) ---
+++ killed by SIGPWR +++
Expected Results: Xfs should reread its config file and do appropriate things.
Since that can be a code-related bug, here is the system info:
Asus P3B-F motherboard, PIII-800, 1G RAM. HDD is IBM DDYS-36950 SCSI,
filesystem is ext3.
BTW, previously I tried to catch the bug on a local machine, and when finally
did it, I did all commands from another one over ssh -- maybe this matters.
Created attachment 61770 [details]
xfs config file
Hmmm... Just tried to do "kill" instead of "killall" for purity, and... It
seems to be a bug in "killall" -- that proggie really sends SIGPWR instead of
SIGUSR1... That is clearly visible if killall is ran as "strace -o
/tmp/killall.log `which killall` -USR1 xfs".
Indeed, this is a bug in killall. I'm reassigning to that package, however
I've been informed that Trond may have resolved this issue already, so I'm
cc'ing him as well.
*** This bug has been marked as a duplicate of 64683 ***