Squid v3.2 bug!

[asaddotcom]

Hi all,

I have FC16 installed and Squid 3.2-13 STABLE on it.. the problem is squid stop working and exited automatically, i don't know why. when i saw in /var/log/messeges it shows:

Jan 18 22:28:02 dotcom abrt[19631]: Saved core dump of pid 11527 (/usr/sbin/squid) to /var/spool/abrt/ccpp-2012-01-18-22:27:43-11527 (1030283264 bytes)
Jan 18 22:28:02 dotcom abrtd: Directory 'ccpp-2012-01-18-22:27:43-11527' creation detected
Jan 18 22:28:02 dotcom abrtd: Size of '/var/spool/abrt' >= 1000 MB, deleting 'ccpp-2012-01-18-14:31:10-9350'
Jan 18 22:28:02 dotcom squid[11525]: Squid Parent: (squid-1) process 11527 exited due to signal 6 with status 0
Jan 18 22:28:04 dotcom abrtd: New problem directory /var/spool/abrt/ccpp-2012-01-18-22:27:43-11527, processing
Jan 18 22:28:05 dotcom squid[11525]: Squid Parent: (squid-1) process 19657 started

does anyone can plz tell what this error log mean into?? and why squid service stop some times automatically and never start again until i don't start it manually??
your help will be appreciated..

Regards

[amturnip]

Some of those messages come from abrt. The abrt tool should help you derive a legible stack dump from the core dump, make a problem report on the Red Hat Bugzilla (https://bugzilla.redhat.com/), and attach the stack trace to the report. If you do not already have a Bugzilla account, you should go get one before using abrt.

[aasami]

I had the same problem with several servers upgraded from older fedoras (as old as F13).

Squid gets restarted it's child processes over and over again due to signal 6 with status 0:

Nov 20 10:47:01 oleg squid[2935]: Squid Parent: (squid-1) process 3026 exited due to signal 6 with status 0
Nov 20 10:47:04 oleg squid[2935]: Squid Parent: (squid-1) process 3045 started

Until it realized that there were too many restarts of child processes an stopped partent too:

Nov 20 11:05:46 oleg squid[2935]: Squid Parent: (squid-1) process 3330 exited due to signal 6 with status 0
Nov 20 11:05:46 oleg squid[2935]: Squid Parent: (squid-1) process 3330 will not be restarted due to repeated, frequent failures
Nov 20 11:05:46 oleg squid[2935]: Exiting due to repeated, frequent failures
Nov 20 11:05:46 oleg systemd[1]: squid.service: main process exited, code=exited, status=1
Nov 20 11:05:46 oleg squid[3333]: squid: ERROR: Could not send signal 15 to process 3330: (3) No such process
Nov 20 11:05:46 oleg systemd[1]: squid.service: control process exited, code=exited status=1
Nov 20 11:05:46 oleg systemd[1]: Unit squid.service entered failed state.

And only when this happend you can find a lead in /var/log/squid/cache.log which reads:

2012/11/20 11:05:42 kid1| assertion failed: comm.cc:1829: "!commHasHalfClosedMonitor(fd)"

According to squid documentation half closed clients is now off by default an as of my expiriences this is the root cause.

Changing half_closed_clients to off in /etc/squid/squid.conf resolved this issue once and for all.

I hope this helps someone. If so, please say thank you to this thread.

Cheers, Aas.