In a recent post to this list I said that that I had seen the entry
"aio_queue_request: WARNING - Queue congestion" in the cache.log of my
newly installed Squid. I was advised not to worry about it.
Nonetheless, I *am* worried about it. I get this message about once per
day and I'm concerned that the occasional warning is a symptom of a deeper
problem. I don't want to have to keep checking Squid to see if it's
running OK.
To recap my config: Squid v2.4S1 (with all posted patches applied) is using
the aufs access type and is using the default 16 threads. The cache itself
is on a solid-state disk, accessed with an UW SCSI interface; the
filesystem is ReiserFS. The overall system is RedHat v7.1 (with all RH
updates applied + v2.4.7 kernel) running on dual Pentium3/550MHz CPUs and
512MB of RAM. This box (and thus Squid) not is not memory contrained in
any way. I am using 16 redirector_children (Adzapper). There aren't any
CPU-intensive jobs running on this machine.
So... why am I getting this warning and what can I do to correct the
problem being warned about?
Thank you.
Received on Tue Jul 24 2001 - 08:24:37 MDT
This archive was generated by hypermail pre-2.1.9 : Tue Dec 09 2003 - 17:01:17 MST