there are two scenario's here to consider

in the normal case when Replicator is running you only need as much savvol space to cover the changes for 2x RTO for the two internal checkpoints

however if the replication has to stop for some reason and you want it to continue or restart without a full base copy then it has to have the space to "survive" for that time.

If you think your replication outages will be less than a day then I would use the daily change rate as a start for the planning

This assume no other user checkpoints

Whats your experience for the other replications you are already using ?

Let me back up a second. How can you tell how much space is being consumed by checkpoints. Either via the GUI or CLI?

sure - the complete space used inside the savvol by all checkpoints is available with nas_fs -s or through the properties of the checkpoint in Celerra Manager

for a individual checkpoint (keep in mind that this is only the blocks used only by that checkpoint - it might also rely on blocks in older checkpoint) you can chart it in Celerra Monitor or since 5.6.39 ns_fs also shows it

Hi,

See http://forums.emc.com/forums/thread.jspa?threadID=135633&tstart=0 .

Gustavo Barreto.

Ok, so I run this command and get this info...

[nasadmin@Celerra ~]$ nas_fs -size ckpt_production6_daily_checkpoint_007
volume: total = 1180000 avail = 1137247 used = 42753 ( 4% ) (sizes in MB)
ckptfs: total = 2959361 avail = 207507 used = 2751854 ( 93% ) (sizes in MB) ( blockcount = 2416640000 ) ckpt_usage_on_savvol: 128MB ( 0% )

Am I correct in understanding that this checkpoint (we retain 15, this is # 7 of 15) is using 4% of its allocated size checkpoint size?

The files system itself is 2.9TB. Am I correct in saying that 1.1TB of space is "set aside" for checkpoints for a 2.9TB file system?

Message was edited by:
DHoffman

the savvol of all Checkpoints of this FS is 660000 MB, 206977 of this is used for the checkpoints all together.

The listet checkpoint (ckpt_production5_daily_checkpoint_007) is using 128 MB of the savvol.

The original FS has total = 2551627 avail = 133715 used = 2417912 ( 95% ) (sizes in MB)

Ok... I think I'm starting to get this..so, using the example above. The volume size is the SAVVOL size for THIS file system or for ALL file systems?

What I'm really trying to find out, is.. how much space is being consumed by the SAVVOL and how much of that is being consumed by any given file system.

So, I ran another command on a different FS.

[nasadmin@Celerra ~]$ nas_fs -size ckpt_production5_daily_checkpoint_007
volume: total = 660000 avail = 453023 used = 206977 ( 31% ) (sizes in MB)
ckptfs: total = 2551627 avail = 133715 used = 2417912 ( 95% ) (sizes in MB) ( blockcount = 1351680000 ) ckpt_usage_on_savvol: 128MB ( 0% )

I dont think this is correct, Rainer


If you look at the data, the savvol is 660000 MB,
the used= on ckptfs is 2417912 MB (nearly 4 times the savvol size) this is the usage of the PFS, not one Checkpoint

I think the 42753 MB and 4% mean the space used by all checkpoints in that savvol
just post the output for next checkpoint - maybe that makes it clearer



yes, 1.1TB is the current savvol size

this isnt the default size though - at some time it must have actually needed that much space and expanded the savvol

its the savvol size that the that file system

each file system has its own savvol - within that savvol all checkpoints for that particular file system allocate space

if you run a nas_fs -info for two checkpoints on different file systems you will see that they are using different device

just go into Celerra Manager under Storage -> Volumes and filter on checkpoint - you'll then see all the savvols on the system with their size (but not how full they are)

actually that number seems to be currently broken - it shows 128MB for any checkpoint

the used= on the ckptfs line seems to list the size for one pariticular checkpoint

Ok, so does this warrant a check by EMC?

what DART version are you using ?

5.6.45-5

I think that depends on the DART code version - on my 5.6.45 system its definitely not the usage of the PFS