Hi! On Fri, 20 Apr 2001, Chris Picton wrote: > I'm runnning into a bit of a problem on our nntpcache machine > > The main nntpcache process seems to stay at about 90% cpu usage (PII 350) > > I am beginning to suspect the disk drives as causing a bottleneck here. > > I have a 35 Gig spool for the news articles mounted on > /usr/local/var/nntpcache. How many users have you got? total & simultaneous.. I don't know what OS you're using. If you're using Linux 2.4, switching to ReiserFS might help for those article directories with multitudes of files. > In nntpcache.config, I have cacheDir set to /usr/local/var/nntpcache/ > > The history files are in /usr/local/var/nntpcache > The cached news articles are in /usr/local/var/nntpcache/news.upstream.net. > > If I were to change the history file to be on a different scsi disk, would > this improve my performance? A different spindle would help, so would RAID.. I can't say I've seen nntpcache chew up tremendous disk on the systems I have run it on. > eg set historyFile to /usr/local/var/nntpcache/history/cache.history, and > mount another scsi on /usr/local/var/nntpcache/history. > > Would there be any other reason for nntpcached to be using so much cpu? NoCEM? If NoCEM's on, definitely turn it off! :-) Cheers, Peter. -- Until I loved, life had no beauty; I did not know I lived until I had loved. (Theodor Korner)
