Network and Storage Protocols

Snapmirror and snapvault

asurendr
6,638 Views

HI,

My Production and Disaster site both have a mix of 32 and 64 bit aggregates. Can we create a snapmirror relationship for a volume in 32 bit aggregte to a volume in 64 bit aggregate. Also, the snapvault source (mirrored volume) and the sanpvault destiantion is on the same filer but a different aggregate.

Since there is mix of 32 an d 64 bit aggregates can we establish mirroring and vaulting among these aggregates. Else, is there one to one relation??

Thanks!!

1 ACCEPTED SOLUTION

scottgelb
6,638 Views

SnapVault is qtree replication so can replicate 32 to 64 or vice-versa.  Volume SnapMirror with ONTAP 8.1 can mirror 32 to 64 bit, but only 8.1+, not 8.0 or 7.3 for example.

View solution in original post

11 REPLIES 11

scottgelb
6,639 Views

SnapVault is qtree replication so can replicate 32 to 64 or vice-versa.  Volume SnapMirror with ONTAP 8.1 can mirror 32 to 64 bit, but only 8.1+, not 8.0 or 7.3 for example.

asurendr
6,595 Views

Scott Thanks!

I created a qtree and a cifs share. Created a share named test_qtree in a qtree named test_qtree (volume: vol_test)

Now i apply the policy to MIRROR AND VAULT it using the protection manager.

prod> Snapmirror status test_vol

Source                                              Destination                                                                                                      State          Lag        Status

hpNetApp1b.hpinc.com:test_vol          rocNetApp1a:TEST_Datasetxforxmirroringxandxvaulting_mirror_hpNetApp1b_te      Snapmirrored   06:10:17   Idle

But when i check the snapvault status. it shows something like this:

DR> snapvault status

rocNetApp1a:/vol/TEST_Datasetxforxmirroringxandxvaulting_mirror_hpNetApp1b_te/-           rocNetApp1a:/vol/TEST_Datasetxforxmirroringxandxvaulting_backup_hpNetApp1b_te/TEST_Datasetxforxmirroringxandxvaulting_hpNetApp1b_test_vol  Snapvaulted    05:40:43   Idle

rocNetApp1a:/vol/TEST_Datasetxforxmirroringxandxvaulting_mirror_hpNetApp1b_te/test_qtree  rocNetApp1a:/vol/TEST_Datasetxforxmirroringxandxvaulting_backup_hpNetApp1b_te/test_qtree                                                   Snapvaulted    05:40:43   Idle

Why is that there are two instances??


scottgelb
6,595 Views

The /- on the source is to specify any data NOT in a qtree and to take all source volume non-qtree data and put in a qtree on the target.  The second instance is your test_qtree.  The cool part is that protection manager is making sure all data is protected in the volume, not just the qtree you created, but any possible data outside of that qtree in the volume.  SnapVault targets are all qtrees, but the source can be non-qtree data in the source volume specified by /-

asurendr
6,595 Views

But nothing resides outside the qtree. Is there any wastage when it vaults even the non qtree stuff. But ideally since there is nothing outside qtree, it doesnt make a difference right??

scottgelb
6,595 Views

Protection manager is just adding everything in the volume it sees (volume outside of qtrees and qtree)..so in this case the first relationship isn't replicating anything... you can try deleting that relationship since not needed.

asurendr
6,595 Views

Scott I have one more thing to ask you. Also I have posted separately

Now, this is the policy that I am applying on the test_vol. I have set the local backup schedule to "none". So this should imply that there should be no snapshots on the test_vol volume right??

But when I hit snap list test_vol, it does show snapshots.  I have set a reserve of 20% default.

Q1. Why is there snapshots when I have selected the primary data local backup schedule to "none"?

Q2. What if the snapreserve is set to 0%. Will there be any snapshots??

here is the snap list output... 20% reserved.

scottgelb
6,595 Views

I posted back via email but don't see it here..might show up delayed... with snapvault there are baseline snapshots...they increment and are used to keep the relationship to the target...so matching snaps incrementing on source and target. With a different retention of snaps on the target but still at least one common snap needed on both source and target.  Having 0% snap reserve will just change reporting...snapshots on the source will show they take active file system space.  an accounting thing. As long as enough room for data and the base snapvault snap then it is ok...but I'd leave a little snap reserve to hold the snaps in my setup.

asurendr
6,595 Views

Yup Scott .. Got it. But what i really wanted is, maybe the image inserted did not show up before, i have inserted it again here.

I have set a policy using the protection manager, on this particular vol_test. Hope you can see the image below. I have set the local backup to "none". . The local backup corresponds to these snapshots right?? or am I wrong??.. If i am right then on what basis are these snapshots being created and what is the default schedule???

Hope these two images below are visible.

scottgelb
6,595 Views

These look like you still have a native snapshot schedule... outside of protection manager.  You can check with "snap sched test_vol" ... then you can change the schedule to "0 0 0" or vol options nosnap on to disable the scheduler in ontap... then manually delete the hourly, nightly snaps.  Leaving the snapvault and snapmirror snapshots.

asurendr
5,781 Views

ok thanks a ton Scott!!

hpNetApp1b> snap sched test_vol

Volume test_vol: 0 2 6@8,12,16,20

So once I remove this schedule and delete those snapshots expect the snapvault and snapmirror. I can then use the local backup schedule for the primary data (in my customised policy) in protection manager to reschedule them again right??

scottgelb
5,781 Views

Correct…but protection manager will manage the schedule

Public