Link to home
Create AccountLog in
Avatar of Relay700
Relay700Flag for United States of America

asked on

Exchange 2010 c: drops below 2GB and message loss????

I was not aware of the 2GB minimum on the c: drive of an Exchange server until I was recently bit by it.  I understand the issue now and added another 25GB to my VM's C: drive.

What I am confused about is that we suffered message on inbound SMTP traffic and I want to understand why to avoid in the future.

We were using Postini's email filtering service at the time.  We are now using McAffee's MXLogic email filtering service.
My understanding of SMTP is that the sending server requires an acknowledgement of the message sent before it will accept it as delivered.  Our mail server obviously was not accepting SMTP traffic during that period so why did Postini's servers believe that the messages were successfully delivered?
During the 15 hour or so period, we completely lost all inbound email and as best we can tell, no senders were notified of any delivery issues.
Postini's failover queue failed to activate and no messages were held for us.

Requests for clarification with both Postini and our vendor, Excel Micro, had little to no information for us nor any suggestions.

Again, i understand what went wrong and took steps to prevent that in the future but I still want to know why there was message loss and could this happen again with MXLogic if something happens to our Exchange server and MXLogic does not correctly identify an issue and fails to queue our incoming messages.

Thanks for any help!
Avatar of joelsplace
joelsplace
Flag of Afghanistan image

MS calls that "backpressure" if I remember correctly.  I've never lost any mail that I know of when my Exchange has stopped working due to backpressure so I would be looking at your spam services.
Avatar of Amit
Use Shadow Redundancy also use separate drive for mail queues. You can change Mail queue path to different drive, so your C drive doesn't go out of disk space.
Avatar of Relay700

ASKER

I looked at the backpressure articles and that appears to be what triggered the halting of SMTP traffic but according the articles, sending servers would have received a rejection notification.  Aparrently this did not happen or Postini's servers simply ignored it.
Either way is a BIG problem.

All of my Exchange databases are located on the D: drive with PLENTY of extra space.
The queue database is something different - not the mailbox databases.
I have to echo above - back pressure is fairly common for those of us who support Exchange and email has never been lost.
Is there something between Exchange and the internet that could have been accepting email from Postini to deliver to Exchange.

Simon.
There is a Checkpoint firewall between the Exchange server and the Internet.
I have the firewall configured to only allow SMTP traffic to the server and only from Postini's email servers. There is no Edge server.
Does the firewall answer the SMTP traffic, scan it and then pass it on?

Simon.
No.  We are not making use of the SMTP services even though it is a Unified Threat Management device.  It would still be scanning for malicious code but that should not be the same as a proper SMTP exchange.
Something has interferred with the SMTP transport somewhere. The Exchange erro when Back Pressure kicks in is specifically designed to stop email from being delivered in such a way that it isn't lost or NDR'd back. So something has not done its job correctly. I haven't heard of Postini reacting badly before.

Simon.
ASKER CERTIFIED SOLUTION
Avatar of Relay700
Relay700
Flag of United States of America image

Link to home
membership
Create a free account to see this answer
Signing up is free and takes 30 seconds. No credit card required.
See answer
Not a solution