HI, Based on the below, how does it actually removed the old data when the cache hit the low_swap value. WIll it remove only selected old data in the cache (can we monitor this) and is there any impact to the users i.e slow performance? Sorry, I' also a newbie in SQUID and reading from the FAQ make me more confused. Really need more experienced person to explain it. Thank you so much.. ************************************* > I have a squid that has been caching for like 10 month. It now have an > amazing size of 4.5 gig. When browsing on the web, it is now very very > slow. I restarted squid with a clean cache. Everything was fine again. This is usually a tell-sign that you have run out o memory. There is a couple of other conditions which could cause this as well, but excessive memory usage is by far the most common. As already indicated the FAQ has good guidance about how much memory Squid uses, in fact a whole chapter about this subject alone.. it is a often misunderstood aspect of Squid. Some of the other common causes to significant slowdown when cache is populated is - Request load on the borderline of what the harddrive(s) can handle. A full cache requires more I/O than an empty cache as removing content from the cache also costs I/O.. - slocate updatedb or other cronjobs trying to index the cache files in vein.. Collecting system statistics with sar or similar tools is a good step to have the required information available to be able to analyze why a system has become slow. It is also worth noting that it isn't uncommon that a freshly restarted Squid runs faster for a while (several minutes) as it takes some time before the client request load is back. By restarting Squid you terminate all currently pending requests, often including large long running downloads.. Regards Henrik __________________________________________________ Do You Yahoo!? Tired of spam? Yahoo! Mail has the best spam protection around http://mail.yahoo.com
