2009-09-17 05:52 AM
I am using sv-smvi to SnapVault snapshots taken with SMVI from one storage controller to another. I am having a hard time understanding what exactly sv-smvi is doing in the background. I setup a SnapVault relationship between the two storage controllers. Through SVMVI, I have two types of "backups" being done for a particular datastore, a daily and an hourly taken every 4 hours. So, I setup an associated SnapVault schedule on the secondary and called them sv-daily and sv-hourly.
Unfortunately, every time I run sv-smvi, it creates two SnapVault snapshots on the secondary, one called sv-daily.x and sv-hourly.x. Now, I would expect it to create the sv-daily.x snapshot on the secondary when I run sv-smvi against my Daily SMVI "backups" and a sv-hourly.x snapshot when I run sv-smvi against my Hourly "backups". However, regardless of which SMVI backup name I specify, sv-smvi always creates both snapshots. Why????
Here is some example output if it helps (netapp1 is the primary, netapp2 is the secondary):
netapp2> snapvault status
Snapvault secondary is ON.
Source Destination State Lag Status
netapp1:/vol/demostore/luns netapp2:/vol/demostore_sv/luns Snapvaulted 08:32:08 Idle
netapp2> snapvault snap sched
create demostore_sv 0@-
create demostore_sv sv-daily 14@-
create demostore_sv sv-hourly 12@-
Does having these separate schedules even make sense?
2010-05-07 06:21 AM
Though this thread is quite old it is yet un-answered, so I give my comment...maybe helpful for other users struggling over it:
I had the same issue, which lead me to the comment posted on Matts page for SV-SMVI:
Although I got not an answer from Matt himself, the next comment was:
"As far as I'm aware, that's still not an option, but Matt said he would work on adding that functionality. Currently, all valid relationships for the source datastore are updated every time sv-smvi is run."
So either you have just one retention, means one snapvault snap sched, or you have to write your own scripts as described in the SMVI BestPractice Guide TR-3737.
With the use of the SDK you can control which schedule should apply.
If the use of rsh is not a problem (security!), you can even more simplify the process by just issue in the script
rsh filer snapvault update -s smvi_<backupname>__recent sv_filer:/vol/dest_vol/dest_qtree
rsh filer snapvault snap create dest_vol sv-hourly
(the snapvault snap create will wait until transfer is completed)
that way you don't have to install the SDK