Thunderstone Support Forums

Hi--

I've noticed that after running certain vortex scripts we're sometimes stuck with leftover instances of vhttpd on the server. This generally happens if our script runs linear searches against a large table.

Thing is, the vhttpds can really stack up if enough searches are run and they soak our 4 CPU system. If we wait long enough they'll eventually go away, but not for a while. Sometimes we have no choice but to kill them by hand.

Any idea what's going on? Is vhttpd processing something in the background perhaps? I ask because when running the "top" command on our box we notice the various vhttpds are only reported at 2-10% of cpu each, but after killing them (let's say there were five total) our server is back to 99% idle time.

Is there a default timeout for vhttpd we should look at?

We're using version 4.04.1067366033 of texis, Linux flavor. Thanks!

The long queries are probably still running after the user hit stop or otherwise went away. Newer versions of vhttpd are better at noticing that.

Your scripts should use reasonable <timeout> setting. The default is 30 seconds.

Thanks Mark. We'll investigate the script timeouts-- they're probably set to 300 seconds, if memory serves.