Hi,
I've found a reproduceable bug in squid 2.3DEVEL3...
I've just increased my cache_dir L1 cache directories back up to 256 from
about 1/4 of that, and run squid -z to recreate the new directories. This
is a Redhat 6.1 Linux system, btw :)
The error message in the cache log is:
1999/12/14 22:19:30| Beginning Validation Procedure
1999/12/14 22:19:30| Completed Validation Procedure
1999/12/14 22:19:30| Validated 15166 Entries
1999/12/14 22:19:30| store_swap_size = 122691k
1999/12/14 22:19:31| storeLateRelease: released 0 objects
1999/12/14 22:19:39| storeAufsOpenDone: (2) No such file or directory
1999/12/14 22:19:39| /var/spool/squid/09/02/00002451
1999/12/14 22:19:39| assertion failed: store_swapin.c:80: "errflag <= 0"
Squid then restarts, with a dirty cache since it crashed out at that point :(
From my squid.conf: cache_dir asyncufs /var/spool/squid 190 256 32
After I removed the swap.state file and let squid totally rebuild it from
afresh, the problem seemed to go away. Arguably a bug that the ordinary
user won't find a problem, but a bug nonetheless :>
Reuben
Received on Tue Dec 14 1999 - 03:29:22 MST
This archive was generated by hypermail pre-2.1.9 : Tue Dec 09 2003 - 16:12:19 MST