forum.purplepixie.org

Hi Dave. During my upgrade from 1.01.9a to Version: 1.02.4/b I received an individual failure alert for each of my nodes, but then I received a single decrease email alert with all nodes listed in the email body. Why did this happen as opposed to me receiving individual decrease email alerts for each node?

Thanks!

Paul

Hi Paul,

Erm first sorry for failures during the update. I really need to sort out the update so it stops the system (and maybe even waits for processes to finish) as it makes changes to the DB and can take a while.

As for the all-in-one it's not a problem.

In days of yore the tester script ran through all the nodes/tests with each node updating relevant alert actions with alert text as applicable. At the very end this script then flushed the alert queues and sent out messages.

Nowadays everyone runs the threaded version that fires up individual test sessions for each node which run through the tests, update the alert action and then flush the queue.

If (as I'm guessing happened) all the nodes ran (or perhaps finished running - were able to update the relevant tables) at the same time then multiple sessions would end up written from a single flush.

HTH,

Dave.

Thanks, Dave. Yeah, when I did the schema update with the experimental script my CPU pegged at 100% for a really long time and any other mysql-based app (e.g. my wiki on the same computer) would fail and report too many mysql connections.

But I still love FreeNATS!

Paul

Hi,

Ooops... Sorry it caused you problems. It added a few new indexes which took a while to build. I'd have thought other mysql databases (or even tables) should have still been working ok if not a bit slow. I'll see if there's some way of throttling the index instructions.

Cheers,

Dave.