Try Out: Windows 7 God Mode

Create a folder in Windows 7 and rename it to

GodMode.{ED7BA470-8E54-465E-825C-99712043E01C}

Then check out its contents. 

Exchange 2007 Mail Submission service stops automatically

Recently I was involved in an issue where Microsoft Exchange Mail submission used to stop silently in day or two on the passive node of the CCR cluster. This used to happen only on one of the Node of the CCR Cluster. There will be no event generated in the logs. This service is responsible for submitting mails from the Mailbox Server to the HUB Server. So in case if the resource fails over to this node then the users will face mail issues. In this case, if the user tries to send a mail, it will stay on their outbox until we manually go and start the service or failback the resources to the other node. While starting the service it used to start without any issue.

I was sure that there is nothing wrong with the server hardware or Operating system or Exchange configuration on this node as there was not a single error getting generated for this issue. I checked the service setting under Service Manager and cross checked with the other node. It was same. I did some goggling but unfortunately couldn't get any answer for this issue. It made me wonder "why am I the only one to face this issue".

After sleeping on this issue for 2 day, I thought of digging in the registry of this node. I went to registry value of HKLM\system\CurrentControlSet\Services\MSExchangeMailSubmission. I starting checking the settings of each registry of this node with the other node. There I found a difference in the values of "FailureActions". On the passive the values were as per the screen Shot in fig 1. While on the passive node it was as per fig 2.

I immediately took the backup of the registry and made the changes on the passive node same as that of the active node. I did some testing by stopping and starting services as well as by doing failover of resources from the active node to the passive node and checking the mailflow. After the testing I failed the resources back to the other node. The services which used to stop earlier in a day or two is now running for over two weeks now.

Note: Please make sure to have a restorable backup of Exchange and registry before performing any changes.