Our newly installed moderately loaded Squid 2.6+diskd has restarted
itself twice for the last day with the same error message in cache.log:
assertion failed: diskd/store_io_diskd.c:384:
"!diskdstate->flags.close_request"
OS FreeBSD 6.2
Squid 2.6.STABLE9 + diskd'd cache_dir
Some random bits from cachemgr's output:
Number of clients accessing cache: 472
Average HTTP requests per minute since start: 6780.2
Requests given to unlinkd: 14321
DISKD Stats:
sent_count: 1229469
recv_count: 1229457
max_away: 73
max_shmuse: 73
open_fail_queue_len: 0
block_queue_len: 146199773
OPS SUCCESS FAIL
open 97702 94851 2848
create 79079 79079 0
close 173899 173893 0
unlink 171021 15639 155382
read 203135 200284 0
write 504633 504629 0
Looks as though diskd's constantly failing to unlink files as the
counter's growing every second. Opens fail less frequently as one can see.
Any suggestions? If any further info is needed, please ask.
