stable + squid 2.5_3

[Mike Hoskins]

I seem to be having a problem, but I'm not sure if it's -stable, Squid, or
neither.  I've been running -stable+Squid on a Dell 4600 for a couple
years now.  All has been well, but after upgrading to Squid 2.5_3 (which
was also done on a couple other, less-loaded, -stable machines without
this issue to date), I occasionally get errors in my cache.log and this AM
Squid died and required a restart.

Here are the lines that get logged to cache.log...

2003/06/09 13:00:15| storeDiskdSend OPEN: (35) Resource temporarily unavailable
2003/06/09 13:00:15| storeDiskdSend: msgsnd: (35) Resource temporarily unavailable

These usually happen in floods (many per second)...

root at eng{squid}# grep OPEN cache.log |wc
   25912  207296 2072960

Before this morning's crash, for example, there were a couple hundred of
these and then finally (maybe the root cause?),

2003/06/09 13:00:15| assertion failed: diskd/store_io_diskd.c:494: "++send_errors < 100"

Restarting Squid seems to make everything "normal" again until it's been
running for awhile.  Similar behavior was seen just after the upgrade
(from previous port version), but I misinterpreted it as a problem with my
old cache...  So I just moved it to cache.old, ran squid -z and now 2.5_3
is running on it's own fresh cache.

I'm going to cvsup to the latest -stable after COB today (PST), and will
monitor the Squid port closely for updates.  If anyone has seen something
similar and has ideas or knows the solution...  TIA.

-mrh

--
From: "Spam Catcher" <spam-catcher at adept.org>
To: spam-catcher at adept.org
Do NOT send email to the address listed above or
you will be added to a blacklist!