I'm not sure which traffic triggered the bug as I have about
10,000-15,000 clients campus-wide.


Regards,


-----Original Message-----
From: Henrik Nordstrom [mailto:[EMAIL PROTECTED] 
Sent: Thursday, March 11, 2004 11:09 PM
To: Michael Vincent Pozon
Cc: [EMAIL PROTECTED]
Subject: RE: [squid-users] Incomplete error pages

On Thu, 11 Mar 2004, Michael Vincent Pozon wrote:

> I think it has something to do with the entry in my cache.log:
> 
> assertion failed: errorpage.c:292: "mem->inmem_hi == 0"
> 
> then Squid dies but restarts itself with DIRTY storage upon
rebuilding.

This is a bug. Any traffic that was ongoing when this happened was 
terminated immediately as Squid restarted.

> then these:
> 
> 2004/03/11 13:10:48| WARNING: newer swaplog entry for dirno 0, fileno
> 00025F38

These are expected on dirty rebuilds.

> 2004/03/11 13:11:40| WARNING: 10 swapin MD5 mismatches

These as well, until the rebuild has finished.

Once the cache has been rebuild only very few of these are expected.

Regards
Henrik

Reply via email to