Bug: very poor admin console performance after 3.6.2 upgrade

Hi folks,

I just upgraded from 3.6.0a to 3.6.2. Well, reinstalled from scratch actually, since the upgrade broke my system. (But that was probably my fault for using an unsupported database which i fixed on reinstall.)

The 3.6.2 admin console seems to suffer from major JavaScript performance bottlenecks compared with 3.6.0a (or it might have been 3.6.0, since the Debian package seemed to think it was 3.6.0). Whenever i click a link, Firefox takes all of 1 CPU core for several seconds, and becomes unresponsive while it is happening.

Is there a workaround? This makes administering openfire very painful.

Thanks in advance,


Could you go more into detail? On which pages does this happen?

I’m not sure how to go into more detail for you. The poor performance happens on every single click on every single page. No other sites (even JavaScript-intensive ones) are affected, so i’m sure it’s not my browser.

I’m not sure how to go into more detail for you.
Firefox version, installed FF extensions, installed Openfire plugins.

so i’m sure it’s not my browser.
Maybe some JavaScript blocking extension?

FF version: Mozilla/5.0 (X11; U; Linux x86_64; en-US; rv: Gecko/2008111318 Ubuntu/8.04 (hardy) Firefox/3.0.4

FF extensions: Netcraft Anti-Phishing Toolbar 1.2, Noscript 1.8.6, Ubuntu Firefox Mods 0.5, Uppity 1.5.1.

OF plugins: Monitoring Service 1.1.1

I use Noscript, but the OF console is permanently allowed. If i deny it JavaScript access, things improve marginally.

If i deny it JavaScript access, things improve marginally.
Then this is not a JavaScript problem. Do you have check CPU load of the server? I had some problems where Openfire took 100% CPU because sent packets in an endless loop. In this case overall response time is slow.

I’ve kept an eye on top as i’ve done some work with the admin console, and there doesn’t seem to be anything significant going wrong. Performance is better now than it was before, even though i’ve re-enabled JavaScript. I think this might have been related to other transient load on the server (it is not a dedicated server).