Re: Filemap limit exceeded?

Hi,

--- Bryan Tonnet <bryan@printman.com.au> wrote:
>
> I'm getting the following error in syslog with
> RedHat 6.0 and Squid
> 2.2.STABLE4
>
> Jan 25 14:28:49 sol squid[19044]: Squid Parent:
> child process 19231 started
> Jan 25 14:29:17 sol (squid): file_map_allocate:
> Exceeded filemap limit
> Jan 25 14:29:17 sol squid[19044]: Squid Parent:
> child process 19231 exited
> with status 1
> Jan 25 14:29:20 sol squid[19044]: Squid Parent:
> child process 19261 started
> Jan 25 14:29:28 sol (squid): file_map_allocate:
> Exceeded filemap limit
> Jan 25 14:29:28 sol squid[19044]: Squid Parent:
> child process 19261 exited
> with status 1
>
> ...and on until squid decides that there are too
> many errors and shuts
> itself down. This is a recent thing in a stable
> configuration that has
> worked well for many months. Has anyone seen this
> problem previously?
>

The size of your cache and mean object size are
specified in your squid.conf in 'cache_dir' and
'store_avg_object_size' directives respectively. By
default, Squid uses 13 Kbytes as the average size.

When allocating the bitmaps, Squid allocates
2 * cache_size / store_avg_object_size
bits for filemaps and you have exceeded this limit.
So, your actual store_avg_object_size is lesser than
what you have set in squid.conf. Make a rough guess
for your average object size by looking into the
object sizes in cache directory and set this to
store_avg_object_size directive in your squid.conf.
This will make squid to allocate the bigger filemap.
Once your squid is running, In the cachemanager tool,
click on the 'General Runtime Information' and find
the mean_object_size. Set this to
store_avg_object_size.

=====
regards
visolve_cache
visolve.com
http://www.v-solve.com
__________________________________________________
Do You Yahoo!?
Talk to your friends online with Yahoo! Messenger.
http://im.yahoo.com
Received on Thu Jan 27 2000 - 03:56:05 MST