I have a new squid binary compiled with delaypools.
This is the only difference between this binary and the old one.
I have 2 squids running on the machine - 1 in accelerator,
1 cacheing.
I want to run the accelerator squid with delaypools.
But, when I cut the accelerator over to the delaypools binary,
I start to see a sudden drop in cache hits and a corresponding
increase in TCP_REFRESH_HIT:DIRECT 's in the backend
cache squid. I can't work out for the life of me what's causing
it. There's nothing special/different about the files that are
apparently being seen as 'stale' as far as I can tell.
What's makes this particularly weird is that I have not changed
the squid.conf file to add the delaypools stanza yet.
If I put the delaypools stanza in, the delaypool works fine,
but I still have the TCP_REFRESH_HIT increase problem.
In each case as soon as I cut back to the original non-delaypools
binary the TCP_REFRESH_HIT's drop off to the original rate.
does this ring any bells with anyone?
cheers,
Mike
Received on Fri Aug 03 2001 - 00:26:32 MDT
This archive was generated by hypermail pre-2.1.9 : Tue Dec 09 2003 - 17:01:27 MST