ctxhx and autofilt are hogging the cpu in Oracle 8.0.5

Hi all,

I'm new to the newsgroup, so please bear with me if this has been asked before.

I have a Solaris 8 machine running Oracle 8.0.5 (patched to 8.0.5.2.0 or 8.0.5.2.1) with ConText 2.4.5.0 (if I remember correctly) at work and it occasionally has a problem, but it's getting worse. The ctxhx process goes wacky and takes a lot of the cpu load. There will often be several of these ctxhx processes in addition to several autofilt processes. (I've seen up to 10 of each!) They don't seem to ever finish and die, so they basically bring the system to its knees. We have killed them manually, but that doesn't always help. More get spawned and the system performance dies again. Will killing them manually (with the "kill" command) cause problems. I've seen up to 10,000 pending requests in the ConText queue. What is the best way to handle this when it happens? Why is this happening? In the past we've deleted the requests from the queue but I've now heard that that is a very bad thing to do. Is that correct? I'm not a DBA so I've looked on the web and in several Oracle books. Nothing gives me any clues. We can't afford a real DBA right now and all of them are probably using 9i anyway.

Anyone's help is very, very much appreciated.

Thanks,

Bob Received on Fri Jan 31 2003 - 07:01:15 CET