I have noticed when doing much the same thing as you with GETs and PURGES to the cache with squidclient that sometime the log will show a PURGE, but squidclient returns and error that the URL was not in the cache. Even when I do a GET, see it in the log, do another GET to see the TCP_HIT to confirm it is in the cache, and a PURGE, I get a MISS error (which I assume means it tried to purge the object but couldn't find it in the cache) I am doing schedule GETs and PURGEs but only a small number of them, and I am using SQUID2/HEAD. I just noticed this behavior yesterday. Nick -----Original Message----- From: Nicole [mailto:nmh@xxxxxxxxxxxxxx] Sent: Friday, September 19, 2008 2:11 PM To: squid-users@xxxxxxxxxxxxxxx Subject: Odd purging problem Hello all Some time ago I had mentioned a problem I was having doing purges on squid. Sometimes the purge will happen, but the tool will complain it never got a reply from a particular server. Othertimes, it complains and no purge happened either. We wrote a small perl tool to contact our squid array and say "delete these URL's." Everything was fine until we switched some to squid3 and now even with squid 2.7. (2.6 works always) The coders feel certain the problem is with squid. When we do a purge, one by one, they always work. When we do a mass purge, we send the deletes through, one url at a time, to each server, then come back around again to delete the next url. It seems to be a problem more when we have to send a quantity of deletes over several. I am curious if perhaps in the newer versions of squid, purges have been given a lower priority, thus they might take a longer time to reply and accept new ones after one has been requested? Anyone have any thoughts on this. Thanks! Nicole -- |\ __ /| (`\ | o_o |__ ) ) // \\ - nmh@xxxxxxxxxxxxxx - Powered by FreeBSD - ------------------------------------------------------ "The term "daemons" is a Judeo-Christian pejorative. Such processes will now be known as "spiritual guides" - Politicaly Correct UNIX Page
