Fatal squid death upon setting cache directory

Fellow Squid users:

        Have run into a strange problem. We've had a pretty successful go
at using Squid 2.2 STABLE 5 on a Solaris 2.6 sparc, until last week, when
this came across in the cache.log:

1999/11/15 18:39:06| Starting Squid Cache version 2.2.STABLE5 for
sparc-sun-solaris2.6...
1999/11/15 18:39:06| Process ID 8762
1999/11/15 18:39:06| With 4096 file descriptors available
1999/11/15 18:39:06| Performing DNS Tests...
1999/11/15 18:39:06| Successful DNS name lookup tests...
1999/11/15 18:39:06| helperOpenServers: Starting 10 'dnsserver' processes
1999/11/15 18:39:06| helperOpenServers: Starting 5 'redirector' processes
1999/11/15 18:39:06| Unlinkd pipe opened on FD 24
1999/11/15 18:39:06| Swap maxSize 15360000 KB, estimated 1181538 objects
1999/11/15 18:39:06| Target number of buckets: 23630
1999/11/15 18:39:06| Using 32768 Store buckets, replacement runs every 2
seconds
1999/11/15 18:39:06| Max Mem size: 262144 KB
1999/11/15 18:39:06| Max Swap size: 15360000 KB
1999/11/15 18:39:06| Rebuilding storage in Cache Dir #0 (DIRTY)
1999/11/15 18:39:06| Set Current Directory to /usr/local/squid/cache
FATAL: Received Bus Error...dying.
1999/11/15 18:39:06| Not currently OK to rewrite swap log.
1999/11/15 18:39:06| storeDirWriteCleanLogs: Operation aborted.
CPU Usage: 0.220 seconds
Maximum Resident Size: 0 KB
Page faults with physical i/o: 0
Memory usage for squid via mallinfo():
        total space in arena: 4437 KB
        Ordinary blocks: 4412 KB 2 blks
        Small blocks: 0 KB 0 blks
        Holding blocks: 432 KB 2 blks
        Free Small blocks: 0 KB
        Free Ordinary blocks: 24 KB
        Total in use: 4844 KB 109%
        Total free: 24 KB 1%

Now, for some reason, we simply can't restart squid.

/usr/local/squid/cache contains an NFS-mounted RAID volume from a NetApp
filer that's worked in this way for well over a year. Perms haven't
changed, and the proxy, as user 'squid', can write and delete to that
volume to it's heart's content. Anyone know what else could be causing
this? We have rebuilt the entire cache (squid -z) twice, but no go.

TIA,

wfms
Received on Mon Nov 15 1999 - 17:04:50 MST