Forum OpenACS Q&A: Re: Monitoring and Controlling Queries

Posted by Steve Manning on

The client isn't an IT professional but is a very savvy finance/mangement bloke and so is led by personal experience and information provided by 'professionals'. As I understand it, the particular report was running a query against a table of sales transactions and it didn't scale well under Tomcat/Postgres. In the first few months it was fine but now after a couple of years it can take several minutes to return and even, I'm told, cause IE to timeout waiting for the page to return. The problem is that they were under the illusion that if they hit the Stop button on the browser it would stop the request and kill the query. Not an unreasonable conclusion if you have no idea how the process occurs, afterall if you kill a report in Access then its dead. So if they think the first request has died and isn't going to come back they press Stop and then request the report again. Now of course there are two queries running so it takes even longer so they hit stop again and so on eventually bringing everything to a fullstop.

I think he's ok now with what we've told him of the way in which OpenACS works but the question got me thinking along these lines. Thankfully I've not needed the solutions but its good to know they're there just in case.