2
Delivery delays / spool backups for the last few months
Question asked by CCC - 3/9/2015 at 9:13 AM
Unanswered
We have seen spool backups happening more often for the last few months, restarting the smarter mail services and local DNS server seems to flush things through.  Seems to happen about once a week, sometimes more.
 
I realize this can indicate many different things (DNS issues, slow RBL, etc).  That said, there are a few similar threads in the forums.
 
 
 
It looks like switching to Google DNS has been a common troubleshooting step, however in production that could wreak havoc with some RBLs so its not a long term solution. 
 
Has anyone else been seeing an increase in this type of issue on SmarterMail v13?  We are running the latest (13.2.5511)
 
I'm not positive if the symptoms coincide with upgrading to 13.x, but that's a possibility.
 
Thanks in advance

12 Replies

Reply to Thread
0
Nathan Replied
Assuming this is for outbound rather than inbound items then several versions back we found that Smartermail could not handle a large spool. We offload all outbound email to Exim (www.exim.org) which can handle high volumes without any apparent issues.
0
Matt Chatterley Replied
We are experiencing the same on two separate installations of SM14 - very frustrating indeed.
0
Adam Lewis Replied
Did anything ever become of any of this? We are on the latest version of SM (15) and have the same issue at least one a week when we get a large influx of spam within a short amount of time. Everything just stops until we stop and start the SmarterMail service.
0
Paul White Replied
What are you system specs?  
My feelings are you are either bottlenecking at the disc or memory.  If your memory usage ( total system) is over 50% , upgrade your memory.  If you are on regular hard disks, upgrade to SSDs.  
WhiteSites.com Blog.whitesites.com
0
Adam Lewis Replied
Off the top of my head I'm not sure. I'll ask my network engineers. That explanation doesn't explain why restarting the Smartermail service would fix the issue and clear out the spool within a few minutes.

Sometimes we've had 15k+ messages clogged up. A restart clears them in 5 minutes.
0
Paul White Replied
Sure it does, When you restart Smartermail all the memory that it was hogging is released. Something many people don't realize is that When your memory usage on Windows goes above 50%, it starts to hit your page file really hard. ( this means windows is now using your hard drive like memory ), If you are on SSD drives the performance hit is not that bad, but on regular hard disks it can bring a system to grinding halt, and if you are on a SAN its even worse as you are dealing the latency of your network.

Another possible cause could be your messages are getting throttled by the big ISPs.
WhiteSites.com Blog.whitesites.com
0
Paul White Replied
My advice is to go nuclear with your firewall. look at the IPs that spam is originating from and starting adding their countries to your firewall. China, India, Russia, Brazil, ext.. I have over 20 countries blocked on my server. Also start blocking sending domains and hosts by their TLD. .top is very common one used for spam. I got very creative with my antispam handling a good year ago.
WhiteSites.com Blog.whitesites.com
0
CCC Replied
We havent had this problem in quite some time.  We moved to Cyren AntiSpam a few weeks later so I suspect the issue could have been related to RBL preformance and or a 3rd party Spam Assassin tool we were using previously.
 
Have you tried looking at your RBLs and/or remote spam assasin servers (if any) making sure that none of them are timing out? 
0
Adam Lewis Replied
We're running Windows Server 2012 with 32 GB RAM with dual Intel Xeon Processors. This server only handles SmarterMail, nothing else.

I still disagree that this is a server issue. The performance of the server isn't affected during these times. The server is still running just fine, the only thing not running fine is SmarterMail. If it is a memory issue, it is related to SmarterMails handling of memory.
0
Adam Lewis Replied
I have not, I will look into that. But I find it unlikely that an RBL timing out would magically start working when we restarted the SM Service.

I do believe that it has something to do with the Spam filtering, because I've noticed that all all of the messages in the Spool will have the status "Spam Check" on them until the restart.

Mmmm.....
0
CCC Replied
In our case, we also had/have a local (Microsoft) DNS server running and a (3rd party) windows Spam Assassin service running on the mail server - those services also got restarted during our troubleshooting steps.

This goes back a year or so - but if memory serves it was the 3rd party spam assassin service that was getting tied up in knots, but restarting the DNS service could have also affected external RBL lookups, especially if there was any negative DNS caching in play.

0
CCC Replied
You might also look at your antivirus exclusions if you haven't already.

Reply to Thread