I hope you noticed that squid3 is still in development... > >On 23.05.06 12:08, Dan Thomson wrote: > >> Whenever I run squid, it seems to run fine until the cache memory > >> becomes full. Once this happens, squid slows down to the point of > >> becoming unusable. > On 5/24/06, Matus UHLAR - fantomas <uhlar@xxxxxxxxxxx> wrote: > >you probably have cache_mem or cache_dir sizes too big. > >check out SQUID FAQ, especially the memory part: > >http://www.squid-cache.org/Doc/FAQ/FAQ-8.html > >and disk cache size recommendation: > >http://www.squid-cache.org/Doc/FAQ/FAQ-4.html#ss4.14 On 24.05.06 15:04, Dan Thomson wrote: > cache_mem 1 GB > The system also has 3 gig of RAM. Note that cache_memory is only used for newly fetched objects, not for those fetched from cache. I'd say this is kind of waste. > Each cache dir is a partition on a SCSI drive (each on separate > drives). The partitions are 60 gig. > cache_dir diskd /var/spool/squid/1 2000 128 256 > cache_dir diskd /var/spool/squid/2 2000 128 256 It's a pretty waste of space, I'd say. for 60GB drives, I'd use size 40000 also, 128 L1 dirs is too much for 2GB cache_dirs, I have 64 on 30GB cache_dirs (but this should not cause any problem) > It also seems as though the network traffic decreases greatly whenever > objects are written to disk cache. this may be an IRQ conflict, or bad drivers for SCSI/net card. > >> I decided to check the storage manager logs to see what was going on > >> when the memory becomes full, and I see a lot of "no valid swapdirs > >> for this object" error messages. what sizes are those objects of? > >> The occurrence of these error messages correlate with a large jump in > >> CPU usage by squid. Is there a known reason for this? May be caused by the fact squid searches for valid cache_dir -- Matus UHLAR - fantomas, uhlar@xxxxxxxxxxx ; http://www.fantomas.sk/ Warning: I wish NOT to receive e-mail advertising to this address. Varovanie: na tuto adresu chcem NEDOSTAVAT akukolvek reklamnu postu. I drive way too fast to worry about cholesterol.
