AWS Thinkbox Discussion Forums

limitgroups gone crazy

Our limits groups seem to have gone crazy… nuke for example has a limit of 246 machines, but the slaves report that ‘max is reached’, even though no slaves are rendering.

In the mongo log we see this:

Mon Jul 1 00:00:59.521 [conn140526410] info DFM::findAll(): extent 1:cef000 was empty, skipping ahead. ns:deadline6db.PermissionGroups
Mon Jul 1 00:01:00.919 [conn140560450] info DFM::findAll(): extent 1:cef000 was empty, skipping ahead. ns:deadline6db.PermissionGroups

Several hundred of these lines

Hey Lazslo, that particular message isn’t really an error, and shouldn’t have anything to do with this issue in particular.

Here’s a couple things I’d recommend. First off, make sure to have the Console panel open in the Monitor to catch any error that might happen in the process.

The first thing I’d like you to try is to use the menu option “Perform House Cleaning”, under Tools (as Super User). This will force a house cleaning, and should theoretically release any unused limits. If there’s any errors in this process, they should show up in the Console panel.

Second, if that doesn’t work, try opening up the Limits panel and right-clicking the nuke limit and selecting “Reset Limit Usage Count”. Again, any errors should be showing up in the Console.

If both of these work without error, we’ll have to do a bit more digging to figure out what’s causing this.

Cheers,
Jon

I ended up restarting the mongo db yesterday, and things seem to work right now.

I also noticed that our ulimit -n setting was reset to 1024 again for some reason, which might have been connected to this.

If i see this error happening again, ill do the checks you mentioned! Also, time to update to the release version…

Privacy | Site terms | Cookie preferences