Hey all squid gurus!
We have a cache to which 6000-7000 clients are
connecting to (2.5STABLE5). On a typical day we are
handling 70-100 requests/sec and 8-11Mbit in network
usage.
The problem I am experiencing is that the machine
starts to swap (performance is still good though). We
have a cache of 90GB, 3GB of memory and cache_mem is
set to 128MB. We are caching quite large files to
speed up frequent downloads (up to 100MB files).
Acccording to the FAQ, this setup should use:
90GB cache = 900MB
cache_mem = +128MB
Add. mem = +20MB
Total recommended: 1048MB x 2 = 2096MB.
I still have 1GB left of memory. No other memory
intense applications are running on this server and
still it swaps out as much as 600-700MB before I
manually restart the cache. It's running on RHEL 3.0.
Any hints are appreciated. If more info is needed,
just let me know. I just don't see the reason to why
it swaps out so much.
--- Memory usage for squid via mallinfo(): Total space in arena: 717313 KB Ordinary blocks: 713918 KB 15096 blks Small blocks: 0 KB 0 blks Holding blocks: 30652 KB 4 blks Free Small blocks: 0 KB Free Ordinary blocks: 3395 KB Total in use: 744570 KB 100% Total free: 3395 KB 0% Total size: 747965 KB --- Thanks! //Martin __________________________________ Do you Yahoo!? Yahoo! Photos: High-quality 4x6 digital prints for 25¢ http://photos.yahoo.com/ph/print_splashReceived on Wed Apr 21 2004 - 00:16:10 MDT
This archive was generated by hypermail pre-2.1.9 : Fri Apr 30 2004 - 12:00:02 MDT