List,
A network hiccup or an unexpected Domain Controller reboot may cause
squid to fill its ntlmauthenticator processes.
When it happen:
FATAL: Too many queued ntlmauthenticator requests (346 on 69)
Squid Cache (Version 2.5.STABLE10): Terminated abnormally.
CPU Usage: 92443.989 seconds = 37626.299 user + 54817.690 sys
Invariably, the next step is a swap corruption:
2005/07/05 08:45:13| assertion failed: store_rebuild.c:79: "store_errors ==
0"
2005/07/05 08:45:16| Starting Squid Cache version 2.5.STABLE10 for
i686-pc-linux-gnu...
This led squid to an permanent "assertion failed" untill I stop Squid,
set swap.state to 1 byte and restart squid.
The question: Can I configure something in Squid to, when
ntlmauthenticators are busy, the cache just... Noop? Or, return an error
screen, or what else... And don't fail miserabily...
Rod.
