Bug 102787 - compile time threading limits are too low for busy caches
compile time threading limits are too low for busy caches
Product: Red Hat Enterprise Linux 2.1
Classification: Red Hat
Component: squid (Show other bugs)
All Linux
medium Severity medium
: ---
: ---
Assigned To: Martin Stransky
: FutureFeature
Depends On:
  Show dependency treegraph
Reported: 2003-08-20 23:32 EDT by Red Hat Production Operations
Modified: 2007-11-30 17:06 EST (History)
0 users

See Also:
Fixed In Version:
Doc Type: Enhancement
Doc Text:
Story Points: ---
Clone Of:
Last Closed: 2006-03-07 05:44:53 EST
Type: ---
Regression: ---
Mount Type: ---
Documentation: ---
Verified Versions:
Category: ---
oVirt Team: ---
RHEL 7.3 requirements from Atomic Host:
Cloudforms Team: ---

Attachments (Terms of Use)

  None (edit)
Description Need Real Name 2003-08-20 23:32:46 EDT
Description of problem:

Seen in squid's cache.log: aio_queue_request: WARNING - Queue congestion

Based on this thread, it appears that the number of threads available to squid
needs to be increased:


The current squid settings use these defaults:

#define NUMTHREADS (Config.cacheSwap.n_configured*16)

which defines 16 threads per cache directory configured in squid.conf.
(reference parse_cachedir() in src/cache_cf.c)

I think that changing the configure line for squid to have
"--with-aio-threads=32" would be a potential starting point.

Either that, or patching the above portion in src/fs/aufs/store_asyncufs.h to be
*32 per cache directory, so that sites already using multiple cache directories
will not be suprised by the overall number of threads decreasing if they have
more than 2 directories setup.
Comment 1 Martin Stransky 2005-11-24 08:30:36 EST
Is it still a problem with the latest RHELs?

Note You need to log in before you can comment on or make changes to this bug.