UNSOLVED
used and available capacity between primary and secondary site differ
Hi, my name is Yves from France,
when issueing nas_pool -size clarata_archive on primary site, the used capacity appears to be 45 384 692 as on the secondary site which is in replication the value is 45 952 832. I investigated on FS and ckpt sizes, no difference between the two sites. By the contrary, nas_volume -info show approximatively the same difference I mean 600 more Go used on the secondary site than on teh primary.
Does someone have a idea of the origin of this difference and how to identify the objects involved in order to return to a good configuration of our NAS.
Excuse my poor English.
Regards
umichklewis_ac7b91
300 Posts
425
0
Posted August 2nd, 2011 08:00
There is a chance it may be root checkpoint usage on the remote side. You can try and add up all the SavVol usage on the remote side. Rainer posted a great script that can get you started - I modified it to count up replication checkpoints:
#! /bin/sh
for ckpt in `nas_fs -query:Type==ckpt -format:"%s " -fields:Name`
do
fs=`/nas/sbin/rootnas_fs -query:Name==$ckpt -fields:BackupOf -format:"%s " `
nr=`/nas/sbin/rootnas_fs -query:Type==ckpt:BackupOf==$fs -fields:Name -format:"%s\n" | wc -l`
echo -n $fs has $nr ckpts savvol " "
/nas/sbin/rootnas_fs -size $ckpt | head -1
done | sort -uSave this as 'savvol.sh' and run it. This will parse your filesystems and report out the number of checkpoints and sizes of checkpoints on your filesystems. This should point out any discrepancies.
Let us know if this helps!
Karl
Thanx Karl and Rainer for this first help; unfortunately we have a harware issue with the CS during the execution of the script... so I am waiting for EMC action to re-establish the correct behaviour of the CS before going on with my investigations in production environnement.
I have notice that the script reports interesting informations in test environnement.
I will let you know asap the result of them and my degree of satisfaction.
Regards
Yves
Rainer_EMC
6 Operator
•
8645 Posts
425
0
Posted August 2nd, 2011 07:00
Hi,
pools arent really used in replication – other than objects like file systems or savvols carve space from a pool.
Most likely your difference is that on one side the savvol’s have a different size.
If you want to troubleshoot list all the dvol’s that are using this pool with nas_volume and compare.
Rainer