Description of problem: squid-2.5.STABLE9-1.FC3.4.x86_64.rpm is broken, child processes continually abort with: Mar 30 14:04:25 mihoshi squid[1119]: Squid Parent: child process 2605 started Mar 30 14:04:56 mihoshi squid[1119]: Squid Parent: child process 2605 exited due to signal 6 Mar 30 14:04:59 mihoshi squid[1119]: Squid Parent: child process 2615 started Mar 30 14:17:18 mihoshi squid[1119]: Squid Parent: child process 2615 exited due to signal 6 Mar 30 14:17:21 mihoshi squid[1119]: Squid Parent: child process 2811 started Mar 30 14:17:23 mihoshi squid[1119]: Squid Parent: child process 2811 exited due to signal 6 Mar 30 14:17:26 mihoshi squid[1119]: Squid Parent: child process 2821 started Mar 30 14:17:29 mihoshi squid[1119]: Squid Parent: child process 2821 exited due to signal 6 Version-Release number of selected component (if applicable): How reproducible: always Steps to Reproduce: 1. 2. 3. Actual results: Expected results: Additional info: downgrading to squid-2.5.STABLE8-1.FC3.1.x86_64.rpm (21-Feb-2005) completely fixes the problem. so squid-2.5.STABLE9-1.FC3.4.x86_64.rpm (29-Mar-2005) is definitely broken. i havent tested i386 yet to determine if its affected or not.
Are you using diskd instead of ufs (cache_dir) ?
yes indeed, diskd was broken somehow in a recent upgrade -- because squid-2.5.STABLE8-1.FC3.1.x86_64.rpm also fails now with diskd. :-/ switching to ufs works... what happened?
Bummer. There goes my theory that it was the -2GB patch that broke it. Does diskd work in 2.5.STABLE7-1.FC3.1?
give me a URL for 2.5.STABLE7-1.FC3.1 :-) fwiw i'm also not running fc3 kernels, i'm running stock unpatched 2.6.8.1 kernel if that has anything to do with it.
This is also true for the starndard x86 Squid RPM. It is fixed in the Fedora Core 4 Test 1 build: http://download.fedora.redhat.com/pub/fedora/linux/core/development/SRPMS/squid -2.5.STABLE9-2.src.rpm This is a know bug, viewing the cache.log will give you more information. It has been reported and fixed here: http://www.mail-archive.com/squid-users@squid-cache.org/msg27210.html The error from my cache.log is as follows: 2005/04/03 10:10:02| Starting Squid Cache version 2.5.STABLE9 for i386-redhat- linux-gnu... 2005/04/03 10:10:02| Process ID 24503 2005/04/03 10:10:02| With 1024 file descriptors available 2005/04/03 10:10:02| DNS Socket created at 0.0.0.0, port 50258, FD 5 2005/04/03 10:10:02| Adding nameserver 192.168.0.1 from /etc/resolv.conf 2005/04/03 10:10:02| Adding nameserver 192.231.203.132 from /etc/resolv.conf 2005/04/03 10:10:02| Adding nameserver 192.231.203.3 from /etc/resolv.conf 2005/04/03 10:10:02| User-Agent logging is disabled. 2005/04/03 10:10:02| Referer logging is disabled. 2005/04/03 10:10:02| Unlinkd pipe opened on FD 10 2005/04/03 10:10:02| Swap maxSize 870400 KB, estimated 66953 objects 2005/04/03 10:10:02| Target number of buckets: 3347 2005/04/03 10:10:02| Using 8192 Store buckets 2005/04/03 10:10:02| Max Mem size: 131072 KB 2005/04/03 10:10:02| Max Swap size: 870400 KB 2005/04/03 10:10:02| Rebuilding storage in /var/spool/squid (DIRTY) 2005/04/03 10:10:02| Using Least Load store dir selection 2005/04/03 10:10:02| Set Current Directory to /var/spool/squid 2005/04/03 10:10:02| Loaded Icons. 2005/04/03 10:10:02| Accepting HTTP connections at 0.0.0.0, port 8080, FD 12. 2005/04/03 10:10:02| Accepting ICP messages at 0.0.0.0, port 3130, FD 14. 2005/04/03 10:10:02| Accepting SNMP messages on port 3401, FD 15. 2005/04/03 10:10:02| Outgoing SNMP messages on port 3401, FD 16. 2005/04/03 10:10:02| WCCP Disabled. 2005/04/03 10:10:02| Ready to serve requests. 2005/04/03 10:10:02| Store rebuilding is 14.5% complete 2005/04/03 10:10:03| Done reading /var/spool/squid swaplog (28188 entries) 2005/04/03 10:10:03| Finished rebuilding storage from disk. 2005/04/03 10:10:03| 28157 Entries scanned 2005/04/03 10:10:03| 0 Invalid entries. 2005/04/03 10:10:03| 0 With invalid flags. 2005/04/03 10:10:03| 28157 Objects loaded. 2005/04/03 10:10:03| 0 Objects expired. 2005/04/03 10:10:03| 0 Objects cancelled. 2005/04/03 10:10:03| 31 Duplicate URLs purged. 2005/04/03 10:10:03| 0 Swapfile clashes avoided. 2005/04/03 10:10:03| Took 0.6 seconds (45292.4 objects/sec). 2005/04/03 10:10:03| Beginning Validation Procedure 2005/04/03 10:10:03| Completed Validation Procedure 2005/04/03 10:10:03| Validated 28126 Entries 2005/04/03 10:10:03| store_swap_size = 782356k 2005/04/03 10:10:03| storeLateRelease: released 0 objects 2005/04/03 10:10:03| assertion failed: stmem.c:93: "current_offset == target_offset" 2005/04/03 10:10:06| Starting Squid Cache version 2.5.STABLE9 for i386-redhat- linux-gnu... 2005/04/03 10:10:06| Process ID 24507 2005/04/03 10:10:06| With 1024 file descriptors available 2005/04/03 10:10:06| DNS Socket created at 0.0.0.0, port 50259, FD 5 2005/04/03 10:10:06| Adding nameserver 192.168.0.1 from /etc/resolv.conf 2005/04/03 10:10:06| Adding nameserver 192.231.203.132 from /etc/resolv.conf 2005/04/03 10:10:06| Adding nameserver 192.231.203.3 from /etc/resolv.conf 2005/04/03 10:10:06| User-Agent logging is disabled. 2005/04/03 10:10:06| Referer logging is disabled. 2005/04/03 10:10:06| Unlinkd pipe opened on FD 10 2005/04/03 10:10:06| Swap maxSize 870400 KB, estimated 66953 objects 2005/04/03 10:10:06| Target number of buckets: 3347 2005/04/03 10:10:06| Using 8192 Store buckets 2005/04/03 10:10:06| Max Mem size: 131072 KB 2005/04/03 10:10:06| Max Swap size: 870400 KB 2005/04/03 10:10:06| Rebuilding storage in /var/spool/squid (DIRTY) 2005/04/03 10:10:06| Using Least Load store dir selection 2005/04/03 10:10:06| Set Current Directory to /var/spool/squid 2005/04/03 10:10:06| Loaded Icons. 2005/04/03 10:10:07| Accepting HTTP connections at 0.0.0.0, port 8080, FD 12. 2005/04/03 10:10:07| Accepting ICP messages at 0.0.0.0, port 3130, FD 14. 2005/04/03 10:10:07| Accepting SNMP messages on port 3401, FD 15. 2005/04/03 10:10:07| Outgoing SNMP messages on port 3401, FD 16. 2005/04/03 10:10:07| WCCP Disabled. 2005/04/03 10:10:07| Ready to serve requests. 2005/04/03 10:10:07| Store rebuilding is 14.5% complete 2005/04/03 10:10:07| Done reading /var/spool/squid swaplog (28188 entries) 2005/04/03 10:10:07| Finished rebuilding storage from disk. 2005/04/03 10:10:07| 28157 Entries scanned 2005/04/03 10:10:07| 0 Invalid entries. 2005/04/03 10:10:07| 0 With invalid flags. 2005/04/03 10:10:07| 28157 Objects loaded. 2005/04/03 10:10:07| 0 Objects expired. 2005/04/03 10:10:07| 0 Objects cancelled. 2005/04/03 10:10:07| 31 Duplicate URLs purged. 2005/04/03 10:10:07| 0 Swapfile clashes avoided. 2005/04/03 10:10:07| Took 0.6 seconds (46167.9 objects/sec). 2005/04/03 10:10:07| Beginning Validation Procedure 2005/04/03 10:10:07| Completed Validation Procedure 2005/04/03 10:10:07| Validated 28126 Entries 2005/04/03 10:10:07| store_swap_size = 782356k 2005/04/03 10:10:07| storeLateRelease: released 0 objects
will this fix be rolled into an FC3 update?
For what it's worth, I'm having major problems with squid-2.5.STABLE9-1.FC3.4. I'm runnning diskd with 3 directories. Downgrading to squid-2.5.STABLE8-1.FC3.1 resolves the problems.
The -2GB patch from upstream breaks diskd. An upstream developer is working on a revised patch. As soon as it's avaliable, I'll build new updates for FC2 (unless we've sent it to Fedora Legacy by then) and FC3. The Squid rpms from FC4T1 predate the -2GB patch, so they're unaffected.
I think this can be closed. squid-2.5.STABLE9-1.FC3.6 fixes this problem for me.
cross fingers, appears to work for me as well.