MailCleaner OpenSource Edition Forum

Have a setup like so

2 Mailcleaners clustered
Primary is also the primary mail receiver
For the main 2 gigs ram, 2.8 quad core CPU (of which it gets 2 of those cores under vmware)
For the secondary 1 gig ram, 2.6 dual core of which it gets 1 core under vmware
hosts loads are usually 20% and quite peppy
Load at the moment is about 0.04 according to TOP, peaked out at 0.12, 163 total processes, nearly a gig free ram and no swap used
approx # of messages a day both process are about 3500.
Each system sits on a different internet connection so there's no overlap
Worked perfectly for about a week and a day after upgrading from the prior release

pages in the admin console all have elements that time out usually with messages like: NOK cannot fetch web service: SOAP-ERROR: Parsing WSDL: Couldn't load from 'http://10.0.2.128:5132/soap/index.php?wsdl' : failed to load external entity "http://10.0.2.128:5132/soap/index.php?wsdl"

Sometimes it'll work and I can see that the filter has hundreds of messages in it that have been there for several hours. Usually I notice that emails get through within a minute of send time even with the filter.

something always times out on the pages, it times out pretty quickly too I've noticed. Tried in Firefox, IE and Chrome.

If the filter's designed to have hundreds or thousands of messages in retry state on the MCB3, that's a new one on me.

Anyway, any idea what I could do to get the web pages to stop timing out and be able to see hy there're tons of messages stuck in the filter?

I appreciate any help

I think I may have discovered the issue. Perhaps not, but this seems to have alleviated most of the problems. It appears that there are a very high number of connections coming out of the mailcleaner units, my guess is several DNS checks per email (reverse, spf lookup? spamcops / all the RBLs, etc...) - my other guess is about 18 connections per email.

It appears that after so many connections the old ones just drop or the new ones don't create, this leads to the filter filling up on waiting to send / waiting to check / etc. Everything starts timing out after 10 seconds on the RBL checks, which I guess would take about 2 minutes per message / fill the queue up like crazy.

My other bet is this has something to do with VMWare as opposed to Mailcleaner, unless there're known TCP/IP issues in mailcleaner, so I will go over to the VMware forums and see if I can't locate that issue.

I moved the primary mail location to the virtual server that had the least network connections going on and my filter has not been above 35 since then