Manual Seeding of database & Log Files

[dixson]

I did a manual seeding of Exchange 2007 on a exchange CCR Cluster. The "last inspected log time" shows the time before the database was dismounted to copy. Hence there is a big queue of backlog log files to be copied to the passive server

What are the precautions to be taken with log files on the active and passive server before manually seeding database files ( Dismounting the active copy and copying the database.edb file to the passive server)? do we need to delete the logs from only the passive server or both the servers

 

[theravager]

If the cluster is connect via high speed links the powershell/gui will reseed itself automatically when you do the reseed. If you have a dedicated nic for ccr then its probibly going to be quicker then you can manually copy the data by a fair bit.

No real precautions to take unless its to resolve some sort of real issue as opposed to something simple like a server update/reboot etc. You could do backup prior this otherwise.

 

[dixson]

My cluster is on WAN. Two diffetent geogriphical locations. I did a manual seeding..
Dismounted the database. Copied the .edb file from ate active node to passive node and mounted the active node again.

Ideally, the replay of the logs should be from the time the database was dismounted.
In my case ( as the get-storagegroupstatus)the logs were copied from the point where the replication had stopped.

we have not deleted and pending logs before we mounted the database....
my question is should we delete the logs on the passive node before we mount the database?

 

[theravager]

I've never done a manual file copy before, pretty sure the command to reseed will delete them itself or instruct you to do it manually.

Regardless of if I am right or not it won't hurt anything, and logic kind of suggest that any logs on there are going to be not used anyway as its not in sync atm, only the logs from your latest backup will be valid anyway, which won't be on there.

 

[dixson]

Thanks....Ideally the log files should not have any impact. however the question was posted, as , after the manual seeding was done, the logs seemed to be copied from the time when replication had stopped....this showed a backlog of logs to be shipped. ideally, it should be zero at the time the the database was mounted after manual seeding...

I got a suggestion on the other forum....before one mounts the database, in such a case, take a backup of "exchange" so all applied logfiles are flushed after the backup takes place....i think the log files that are already applied to the databse will grt flushed, i have my doubts if the get lfushed if they are not applied to passive database too!!!