That means that a lot of disk IO is being done and you've run out of threads
and work queues.
You'll want to start by analysing your IO patterns and seeing if you're maxing
out your disk/controller/etc. If you are then you may find some COSS/AUFS
tuning to optimise things, but its difficult to tell what from where I'm sitting.
Adrian
On Wed, Jun 18, 2008, liwei wrote:
> Hi,all
>
> I got a lot of messages when I porting squid2.5 to squid2.6
> with coss,who know why? Thanks!
> ========================================================================
> ===============================================
> 2008/06/18 11:45:14| squidaio_queue_request: Async request queue growing
> uncontrollably!
> 2008/06/18 11:45:14| squidaio_queue_request: Syncing pending I/O
> operations.. (blocking)
> 2008/06/18 11:45:14| squidaio_queue_request: Synced
> 2008/06/18 11:45:33| squidaio_queue_request: WARNING - Disk I/O
> overloading
> 2008/06/18 11:45:34| squidaio_queue_request: Async request queue growing
> uncontrollably!
> 2008/06/18 11:45:34| squidaio_queue_request: Syncing pending I/O
> operations.. (blocking)
> 2008/06/18 11:45:34| squidaio_queue_request: Synced
> 2008/06/18 11:45:53| squidaio_queue_request: WARNING - Disk I/O
> overloading
> 2008/06/18 11:46:05| squidaio_queue_request: Async request queue growing
> uncontrollably!
> 2008/06/18 11:46:05| squidaio_queue_request: Syncing pending I/O
> operations.. (blocking)
> 2008/06/18 11:46:05| squidaio_queue_request: Synced
-- - Xenion - http://www.xenion.com.au/ - VPS Hosting - Commercial Squid Support - - $25/pm entry-level VPSes w/ capped bandwidth charges available in WA -Received on Wed Jun 18 2008 - 07:15:52 MDT
This archive was generated by hypermail 2.2.0 : Wed Jun 18 2008 - 12:00:03 MDT