Unexpected DAG mailbox database failover

Question

Good evening
I am running a three node Exchange 2013 server environment in a stretched DAG configuration. I have been experiencing a problem recently whereby failover is occurring unexpectedly. Below is a brief description of the environment.
Site 1 (Head Office)
2 x Exchange 2013 CAS/MBX (dual role servers, lets call them A and B)
Both are members of the only DAG we have
They both have copies of the only mailbox database we have
Server A has activation preference 1
Server B has activation preference 2
Site 2 (Datacenter)
1 x Exchange 2013 CAS/MBX (dual role server, lets call it C)
C is a member of the same DAG as above
It hosts a copy of the same mailbox database as above
Server C has activation preference 3 (although I have configured it to never automatically activate a database copy)
Server C will only be actively used in a disaster scenario where we lose both A and B in the head office for whatever reason
So recently I have noticed that when the mailbox database is mounted and active on server A after a period of not more than an hour or two, the database is automatically moved to server B. This has happened 3 times now.
I find these informational alerts in the event viewer on server A.
Event ID 2136
Log Name: Application
Source: MSExchangeRepl
Date: 14/11/2014 16:01:03
Event ID: 2136
Task Category: Service
Level: Warning
Keywords: Classic
User: N/A
Computer: ServerA.domain.internal
Description:
Unable to communicate with the Microsoft Exchange Information Store service to coordinate log truncation for database 'Mailbox Database 1\ServerA' due to an RPC communication failure. Error: 3355381764
Extended error: Failed to open a log truncation context to source server 'ServerA.domain.internal'. Hresult: 0xc7ff1004. Error: Error returned from an ESE function call (-1305).
Event Xml:
<Event xmlns="http://schemas.microsoft.com/win/2004/08/events/event">
<System>
<Provider Name="MSExchangeRepl" />
<EventID Qualifiers="32772">2136</EventID>
<Level>3</Level>
<Task>1</Task>
<Keywords>0x80000000000000</Keywords>
<TimeCreated SystemTime="2014-11-14T16:01:03.000000000Z" />
<EventRecordID>4381652</EventRecordID>
<Channel>Application</Channel>
<Computer>ServerA.domain.internal</Computer>
<Security />
</System>
<EventData>
<Data>Mailbox Database 1\ServerA</Data>
<Data>3355381764</Data>
<Data>Failed to open a log truncation context to source server 'ServerB.domain.internal'. Hresult: 0xc7ff1004. Error: Error returned from an ESE function call (-1305).
</Data>
</EventData>
</Event>
Event ID  3169
Log Name: Application
Source: MSExchangeRepl
Date: 14/11/2014 16:01:04
Event ID: 3169
Task Category: Service
Level: Information
Keywords: Classic
User: N/A
Computer: ServerA.domain.internal
Description:
(Active Manager) Database Mailbox Database 1 was successfully moved from ServerA.domain.internal to ServerB.domain.internal. Move comment: Managed availability system failover initiated by Responder=RpsDeepTestPSProxyFailover Component=RPS.
Event Xml:
<Event xmlns="http://schemas.microsoft.com/win/2004/08/events/event">
<System>
<Provider Name="MSExchangeRepl" />
<EventID Qualifiers="16388">3169</EventID>
<Level>4</Level>
<Task>1</Task>
<Keywords>0x80000000000000</Keywords>
<TimeCreated SystemTime="2014-11-14T16:01:04.000000000Z" />
<EventRecordID>4381655</EventRecordID>
<Channel>Application</Channel>
<Computer>ServerA.domain.internal</Computer>
<Security />
</System>
<EventData>
<Data>Mailbox Database 1</Data>
<Data>ServerA.domain.internal</Data>
<Data>ServerB.domain.internal</Data>
<Data>Managed availability system failover initiated by Responder=RpsDeepTestPSProxyFailover Component=RPS.</Data>
</EventData>
</Event>
Does anyone have any ideas as to why this might be happening or whether there are any additional troubleshooting steps I can take?
Regards

Answer 1

Hi George,
Agree with Jared,  running CollectOverMetrics.ps1 script is a good choice to troubleshoot the issue.
Except this method, you can also refer to the following article to troubleshoot the issue with other useful information, like mailbox database replication status information, HighAvailability channel event logs..
http://technet.microsoft.com/en-us/library/dd351258(v=exchg.150).aspx
Best regards,
Niko Cheng
TechNet Community Support