Unsolved

This post is more than 5 years old

19 Posts

6532

December 27th, 2012 05:00

Removing checkpoints

We have had write performance problems that were so bad, we had to move a major email archvie storage application from NAS (file level) back to a Virtual Server (block level).  We have read that others have this problem due to checkpoints.  We have 42 checkpoints (3 per day for 2 weeks) for each of 2 of our 5 file systems.  We have daily Backup Exec backups running, so we are going to try removing all checkpoints and see if performance changes.  I have already paused the 2 scheduled checkpoint events for fileserver1 and fileserver2 (our CIFS servers and related filesystems).  So, I want to begin deleting checkpoints now.  Is there a best practice for this?  Should I start with the oldest checkpoint and do the one at a time?  Should I remove all of the checkpoints associated with a particular filesystem before deleting the second set of checkpoints?  Will there be an impact on my SAN while removing checkpoints?  I am going to track performance testing as I delete these checkpoints.  The theory is that due to the number of checkpoints and the type of storage it resides on, that it adds up to bad write performance.  Anyway, any advice would be helpful.  I want to try and get these removed today if possible.

Thank you.

11 Legend

 • 

20.4K Posts

 • 

87.4K Points

December 27th, 2012 05:00

i thought this performance issue was resolved with the latest File OE version ?  I always delete checkpoints from oldest to newest to older, not sure if it makes any different but that's how i do it. I don't think it matters one checkpoint or many checkpoints, file system is doing COFW so if there was any issue related to that it would happen even with just one checkpoint.

11 Legend

 • 

20.4K Posts

 • 

87.4K Points

December 27th, 2012 06:00

don't remember, looked through some of the release notes

19 Posts

December 27th, 2012 06:00

Which File OE version would that be exactly?

19 Posts

December 27th, 2012 10:00

Deleting checkpoints now and it seem to be going well.  Once all are gone, will I see my savvol space per file system change right away or is there some clean task that must be done?  That is another reason I am removing these is to get needed white space back from bloated savvol space.  I am only using 3% of 5.6TB of savvol space at this time.  I hope to get that back and maybe start up checkpoints again in the future.

11 Legend

 • 

20.4K Posts

 • 

87.4K Points

December 27th, 2012 10:00

savvol will be released as soon as all checkpoints are deleted (including any checkpoints used for replication)

19 Posts

December 27th, 2012 10:00

Ok, so, even if I delete the 42 checkpoints I created via a schedule, my 5.6TB of savvol space will be there until i remove the replication checkpoints?  yikes!  I am replicating this filesystem to a secondary site.  How do i clear out the replication checkpoints and then let them start again?  Do I have to pause or stop replication?  Will I have to begin replication anew?

11 Legend

 • 

20.4K Posts

 • 

87.4K Points

December 27th, 2012 11:00

Nope, replicator needs internal checkpoints ..changes between two replicator checkpoints determine what has changed/new and needs to be replicated .

11 Legend

 • 

20.4K Posts

 • 

87.4K Points

December 27th, 2012 11:00

delete replication and start fresh

19 Posts

December 27th, 2012 11:00

Then will having to use checkpoints for replication to work also cause a hit in performance?

19 Posts

December 27th, 2012 11:00

I cant pause replication, delete the existing checkpoints and un-pause replication?

11 Legend

 • 

20.4K Posts

 • 

87.4K Points

December 27th, 2012 11:00

potentially, you probably have seen this discussion by now

https://community.emc.com/thread/116347?start=0&tstart=0

1 Rookie

 • 

17 Posts

November 22nd, 2013 11:00

What is the impact of deleting the checkpoints and starting replicator fresh? is it a full copy resynch?? Or via the checkpoints an incremental only?

11 Legend

 • 

20.4K Posts

 • 

87.4K Points

November 22nd, 2013 12:00

if you delete replication, that in turn deletes two "hidden" checkpoints that are used specifically for replication ..and that means full resync.

6 Operator

 • 

8.6K Posts

November 22nd, 2013 13:00

Unless you have a synchronized user checkpoint there

6 Operator

 • 

8.6K Posts

November 22nd, 2013 13:00

In that context - no

0 events found

No Events found!

Top