You won't like the answer, but our 7.1.00.002 server has experienced random 
hangs of the MAPI email system since it went live in May 2008, with an Incident 
opened in June 2008 that was NEVER RESOLVED.  Also on Windows 2003 x64, with 
Outlook 2007 SP1 connecting to Exchange 2007, with archiving set up to keep the 
number of sent emails in Exchange down to a month's worth (25,000 to 40,000 
messages).  A nightly escalation moves all Sent mail to an archive folder, so 
there is never more than one day's mail in the Messages form.  Outgoing and 
incoming mail interval is 3 minutes, which usually keeps the queue below 60 
messages at any given interval.

Over the years since this system went live, support has shipped me everything 
they could think of in the way of modified or instrumented email engines (up to 
7.1.00.004), and nothing that they came up with prevented the three types of 
errors:
1. The aremaild.exe process hangs at over 1200 handles with no errors generated 
in messages or logs of any kind (normal is 700 to 800 handles idle, 1000 to 
1300 while processing)
2. A java error appears in the Event Log and the service is hung
3. A error is logged in an hs_err_pid####.log file in the \Office12 folder, and 
the service is hung

In most cases the Windows service is hung and cannot be manually stopped in a 
normal fashion - it will stop eventually, but issues a message about not 
responding. Rarely, you might have to kill the actual process from the Task 
Manager. Also rarely, the service is still running but MAPI processing has 
stopped (our patch 4 build writes a mapi.log continuously) and you can restart 
the service more or less normally (does not hang while trying to stop).

I have had problems with the AREmail engine on Windows with MAPI ever since 
they switched it to java, it's just dramatically worse in 7.1.  We just live 
with it, keeping a query open in a User Tool 24 hours a day where we can see it 
for Email messages where Send = Yes, refreshing every 5 minutes, and when the 
screen fills (= over 100 messages) we restart the service.  A schedule task 
runs a batch job three times a day, 5:45 AM, PM, and 11:45 PM that does a net 
stop and a net start of the service.  I disabled a fourth one at 11:45 AM 
because whenever it fired with a lot of messages in the from waiting to go out, 
it often hung within the next 15 minutes.

During testing with ARS 7.5 on Windows 2008 R2 we saw even worse problems with 
MAPI - the service would not process mail unless you had a console open, logged 
in under the account that ran the email service, and you had started the 
service from there!!  We plan to use SMTP and POP for 7.6.0x; it seems to work 
reliably enough, but then so did 7.1 MAPI back in 2008, until it was put under 
a normal production load.

Christopher Strauss, Ph.D.
Call Tracking Administration Manager
University of North Texas Computing & IT Center
http://itsm.unt.edu/
From: Action Request System discussion list(ARSList) 
[mailto:[email protected]] On Behalf Of Wangler, Dan
Sent: Tuesday, March 29, 2011 8:44 AM
To: [email protected]
Subject: Email Issue

**
Listeners

I have an annoying issue with my email engine.  We send out approx 20K 
email/day from our ITSM 7.1 system.  We started out almost 2 years ago 
interfacing with Exchange but, due to the large number of email that we are 
generating, we moved to an SMTP server (Barracuda appliance).  We were on the 
Barrcuda for many months without incident.  Then suddenly, the email engine, 
while still sending out many email, started randomly hanging.  It takes a 
recycle of the email engine process to restore service.  Inbound email from 
Exchange continues to come in.

The Email Engine may work for hours before hanging up or may be as little as 10 
minutes depending on the amount of email in the queue.  I may hang up during 
peak times but it also hangs during non-peak times.

We switch back to the Exchange Server, but, while it is more reliable than the 
SMTP server, it also hangs.

Nothing in the Email Engine logs (I set the java logging to FINER 
(com.bmc.arsys.emaildaemon.level = FINER) indicate anything more than normal 
email coming and going.

We had SendEmailSetSize = 500.  Thinking we were overwhelming the mail server, 
we set it down to 50.   But, either way, it continues to hang.

Has anyone experienced issues with the Email Engine like this?  Are there any 
other logs I can look at that may help determine what the Email Engine is 
hanging on?  Can anyone offer any suggestion that my help resolve this?

We are running on ARServer 7.1 patch 9.  We are using JRE:1.6.0_10 on a Windows 
2003 server and MSSQL 2005 Server.

Thanks and regards

[cid:[email protected]]

Dan Wangler, MHS Remedy Support Team
Science Applications International Corporation
MHS Cyberinfrastructure Services (MCiS)
300 Convent St, Ste 1800, San Antonio, TX 78205
210-338-3162 * 972-979-1970 * 
[email protected]<mailto:[email protected]>





_attend WWRUG11 www.wwrug.com ARSlist: "Where the Answers Are"_

_______________________________________________________________________________
UNSUBSCRIBE or access ARSlist Archives at www.arslist.org
attend wwrug11 www.wwrug.com ARSList: "Where the Answers Are"

<<inline: image001.jpg>>

Reply via email to