Monday, January 23, 2017

Get-MoveRequest Queued "Job is waiting for resource reservation"

The other day I was trying to move a 5 MB mailbox from one server to another (both running Exchange 2013) when I noticed the move request was stuck on Queued for a long time. The first thing I did was checking the statistics of the move request by running the following cmdlet:
Get-MoveRequest | Get-MoveRequestStatistics | FL

In the stats I could see the following warning/error:
Job is waiting for resource reservation. MRS will continue trying to pick up this request. Details: Resource reservation failed for 'LocalServer/ServerRead' (Processor): load ratio 4.2, load state 'Overloaded', metric 64. This resource is currently unhealthy.

This would change intermittently to the following warning/error:
Message: Resource 'Processor' is unhealthy and shouldn't be accessed.

This could also be seen in the Application Event Log of the server:
Log Name:      Application
Source:        MSExchange Mailbox Replication
Date:          1/19/2017 10:25:09 AM
Event ID:      1121
Task Category: Request
Level:         Error
Description:   The Microsoft Exchange Mailbox Replication service was unable to process a request due to an unexpected error. Request GUID: 'e2f8d856-f258-4cab-a1d1-dde19df2a000' Database GUID: '3479f71f-df65-48ff-a80d-9379495b6aac' Error: Resource 'Processor' is unhealthy and shouldn't be accessed.

The cause is self-explanatory: high CPU usage. When I checked the server’s CPU, this was indeed the case:

The possible workarounds are to stop any CPU-intensive processes (if there are any that can be stopped), investigate what is causing the high CPU and fix it, or wait for the CPU usage to come down at which point the move request will resume automatically.

You may also see similar warnings for other resources such as disk IOPS for example:
Resource reservation failed for 'MdbWrite(“database_name”)' (MdbLatency(“database_name”)): load ratio -1, load state 'Unknown', metric (null).
Resource reservation failed for 'Mailbox Database/MdbWrite' (CiAgeOfLastNotification(“database_name”)): load ratio X.XXXXXXX, load state 'Critical', metric 2147483647. This resource is currently unhealthy.

No comments:

Post a Comment