100% processor usage

[DTower5]

Hi,

Just wanted to bounce some stuff off and see if there is something really wrong.

Every so often, my AMS process will jump and stay at 99% of processor time, during this time - I cannot access the webmail, but can access the other services. The process is clearly leaking at this point, as it will continue to build mem usage, about 2K at a time. If I kill and restart the service, it will come back up as normal until the next iteration of this problem. This seems to happen more often when the server is busy (afternoons)

I did test to see if the problem only happened in service mode, and it does not, the same problem occurs when running as an application.

My webmail log doesn't really show anything out of the ordinary, other than the break in time during the hang

Code: Select all

Tue, 25 May 2010 16:47:50 -> 75.175.214.252 -> Success: Action=[Close Connection]
Tue, 25 May 2010 16:47:54 -> 8.6.95.143 -> Success: Action=[Accept Connection], Details=[Port 443: Implicit SSL]
Tue, 25 May 2010 16:47:54 -> 8.6.95.143 -> Success: Action=[HTTP Request], Details=[GET \_index - OK]
Tue, 25 May 2010 16:49:24 -> 8.6.95.143 -> Success: Action=[Close Connection]
Tue, 25 May 2010 17:13:41 -> Success: Action=[Init Listening Ports], Details=[Listening on port 8080 and 443.]
Tue, 25 May 2010 17:13:54 -> 63.230.150.115 -> Success: Action=[Accept Connection], Details=[Port 443: Implicit SS

Any ideas? Thanks!

Relevant Info

P4 3.1 GHZ
1.25 GB RAM
300 GB SATA HD
AMS 2.70

[Code Crafters]

Without any logs to help support what you're saying, it's hard to tell if there is a bug or what the cause of that would be. I can only suggest maybe getting some more memory especially if you are running Vista as this runs very poorly with less than 2GB RAM. Other than that, maybe some other users have some ideas of a cause. If I can replicate the problem we can fix it so let me know if you find the trigger of the issue.

[DTower5]

I also just tried reinstalling the program, and that did nothing - it seems to be getting worse, but at this point I don't want to redo the box -

I was playing around with using a second machine to do the webmail - but anytime I try point a user account on the backup machine to the remote drive with account info - it says it already exists (it does) and that any config settings would be overwritten - is there a way around this?

[Code Crafters]

If the logs don't show any abusive activity, check the main admin dialog status page to check for excessive connections running. Maybe increase the allowed number of WebMail simultaneous connections allowed from the WebMail settings too. Moving accounts to another drive probably won't make any difference to this issue as it's more a WebMail service issue which should show up in the WebMail logs really. However, if you do move a user path to another drive it will overwrite any files already there with the settings that are loaded into memory already. Also, mails will need to be moved to the new remote drive; ideally just copy the old user path folder to the new location but again I don't think this is related to the issue you have.

[centralusa]

one thing that we have seen cause 100% cpu is anti virus software checking on a busy server