Forum Discussion
Autoreseed, now what?
Have had a disk failure in a four server Exchange SE DAG with autoreseed enabled. New disk inserted, but now what? What I can google and AI myself to is something like this:
- Bring the new disk online
Remove the broken mount point by deleting the mount point folder that does not lead anywhere
Create a New Simple Volume and mount it in an empty NTFS folder
- Format it as per our standard, ReFS 64K and label to our standard (same as the old one)
Does the experts agree that this is all there is to it?
Many thanks!
4 Replies
AutoReseed and the Disk Reclaimer work together to automate this process using one or more spare disks that you pre-provisioned. Here's how it works:
The Microsoft Exchange Replication service periodically scans for any database copies that have a status of FailedAndSuspended. If it finds all database copies on a volume configured for AutoReseed in a FailedAndSuspended state for 15 consecutive minutes, the AutoReseed workflow begins:
- AutoReseed tries to resume the failed and suspended copies up to 3 times, with a 5-minute wait between each try. Note that sometimes after a FailedAndSuspended database copy is resumed, the copy remains in a Failed state. To handle those cases, AutoReseed automatically suspends a Failed database copy for 10 consecutive minutes while its workflow is running. If the suspend and resume actions don't result in a healthy copy, the workflow continues.
- When it finds a copy with FailedAndSuspended status, it verifies that a spare disk is available, and that the database and its log files are on the same volume and in the right locations using the required naming conventions. AutoReseed also requires you to use a specific storage structure that uses mount points to pre-map disk volumes (including your spares) to databases. If these checks pass successfully, the Disk Reclaimer allocates, remaps, and formats a pre-provisioned spare disk. The Disk Reclaimer attempts to format spare disks at different intervals (but only once per day), depending on the state of the disk and database copies. For a disk to be automatically formatted, the volume must have a mount point in the root volumes path (by default, C:\ExchangeVolumes) and must not have any mount points in the database volumes path (by default, C:\ExchangeDatabases).
- AutoReseed then attempts to assign the spare volume up to five times, with a 1-hour wait between each try.
- After a spare is assigned, AutoReseed performs an InPlaceSeed operation with the SafeDeleteExistingFiles switch. All databases on the failed disk are reseeded on the spare using the active copy as the seeding source.
- After the seeding operation is completed, the Exchange Replication service verifies that the new copies are healthy.
If all retries in the above workflow are exhausted, the workflow stops. If after 3 days the database copy is still FailedAndSuspended, the workflow starts over from Step 1. This behavior is intentional because it can take some time to replace failed hardware.
Hope this helps!
Thank you very much for your explanation!
So in a post event, when this has happend, is that action to take to remove the old mount point and create a new one?
"For a disk to be automatically formatted, the volume must have a mount point in the root volumes path (by default, C:\ExchangeVolumes) and must not have any mount points in the database volumes path (by default, C:\ExchangeDatabases) "
https://learn.microsoft.com/en-us/collections/g2ggf3tz1npmqd?terms=unleash&sharingId=5C206E23C6D1789F
Not sure what this link gives?
