Hi there, newbie here so please go gentle.

Installed 5.3.3-B1879, using base MSSQL CE, on a Windows Server 2003 (Standard Edition with SP2) Thursday and then set it up properly Saturday. 1 x domain, 9 x accounts and one distribution list.

As regards server settings, I don't have a lot set - under performance

Caching is enabled at 60-70 and am getting hit rates @ 90-50%.
Threading Command Threads and Delivery Threads both set to 20 with synch task threads at 15.
I have promoted the Working Thread Priority to Above Normal
Did not set up indexing as not using web front end.


All appears to work ok internal and external mail behaving as expected, except for 2 occasions (Sunday ~17:00 & Today ~14:30) when the server stopped responding to IMAP & POP requests and the clients (Outlook 2007 + 1 x remote iPhone pop) all went offline.

I've done the basic faultfinding…
IP comms between clients and server OK
telnet to server port 25 gives smtp response
hmailserver admin tool OK.

Based on comments in the forum I've restarted hmailserver service from Windows Service Admin tool, which appears to have got things up and running on both occasions.

I've attached the error logs, full logs available from dropbox…

http://dl.dropbox.com/u/12297430/hmail_logs.zip

Saturday's drop appears to have coincided with 1 or 2 problematic messages (available from dropbox)

Can't see anything on Sunday but admit to being a bit lost.

Any and all advice Appreciated

You have a bit of a mess there...

I can go days without anything written to error logs, in fact I get a single error log line, maybe three or four days a month in total.

What AV do you have on the system?
What Router are you using?

Looks like even though only 10 accounts, that the server is pretty busy. You should look at a real database very soon.

Also, the IOP errors in the error logs is a known issue with 5.3.3
If possible look at upgrading to 5.4

Most of the errors were about broken communication, which makes me think AV or something netork related

Thanks for the response (I did think it was a bit of mess, but as this is not my day job I wasn't sure if it was normal )

Just been on remotely via Teamviewer to check and although there hasn't been an outage today, got to feel there's one coming - the connection was v. slow, so I checked task manager and the CPU is bouncing along at 98-99%

Don't think there's any issues with the router - it's a cisco/linksys twin lan port running off 2 x adsl lines. It's log is clear and network response (other than email) is reported as good.

Did find that current AV is out of date so isn't actually working - it was Kapersky, and is still running, throwing up the occasional nag screen I guess uninstalling this might free up a bit of cpu but not likely to make a significant impact is it? - There's a new server ordered (to replace the original system that this is a replacement for) that will come with a new AV contract etc, so my mate is not keen to renew anything now. Issues is that I don't know when it's going to be available so I need to keep this up for a week or so. Could I use Immunet in the interim? Would that help with volume of errors?

If I can get in front of it tonight, or free up some CPU time I'll try the 5.4 update.

Looking at todays logs and the mails that are actually coming in, there does appear to be vast volumes of spam. This isn't my day job, but is that normal? I didn't enable any of the AV/SPAM features on the server because I thought it would suck up resource, but thinking about it, might it help the cpu usage if I was to enable some of the filtering features?

Probably bad form to post to yourself, and may be tempting fate since it's less than a day since the changes, but... Upgraded to 5.4-B1931 and changed some set-up and everything seems to have calmed down a lot.

No ERROR logs at all Server CPU now @ 25%

What I did...

Uninstalled expired Kapersky AV
Installed ClamWin
Stopped the service, backed up the hmailserver folders, upgraded to 5.4-B1931
Changed the following settings...

Protocols - set max number of sessions to 10 for SMTP, 4 for POP and 20 for IMAP - we had been getting a lot of external login attempts, and there's only a couple of people who should be looking for mail from outside.
Anti-Spam - enabled the default SURBL server, turned on Grey listing (need to do some reading to find out what that actually is).
Anti-Virus - Enabled ClamWin (& tested OK)
Advanced-Auto-Ban - Enabled with default settings (Now have 10 Banned ranges)
Advanced IP-Ranges - Disabled External-External Delivery for all ranges, required smtp authorisation for everything except external-local from the internet & router
Advanced-Performance - Wound up the cache values.

Tks

If possible use ClamAV instead of ClamWin. Clamwin is only meant for small use home type servers, not production servers. Clamwin uses a lot of CPU resources especially on busy mailservers.