UPDATE: For ONTAP 9.0 and later see Creating
SVM-DR Relationship post 7 to C Cutover of DR and Prod
If you’ve got big volumes to transition from 7 to C, and want SVM-DR functionality almost on the button post 7 to C cutover (i.e. it’s not acceptable to have to wait for the entire SVM to SnapMirror across to DR), it’s possible to pre-stage the SVM-DR relation. Here’s an example to show how:
If you’ve got big volumes to transition from 7 to C, and want SVM-DR functionality almost on the button post 7 to C cutover (i.e. it’s not acceptable to have to wait for the entire SVM to SnapMirror across to DR), it’s possible to pre-stage the SVM-DR relation. Here’s an example to show how:
Note: cDOT version here is 8.3.2. 7MTT version
is 2.3.1.
1)
Create the initial baseline data copy from 7-Mode to cDOT.
Image: 7-Mode volume baselined to cDOT
1b) Extra Belt and Braces Step
Create a snapshot on the 7-Mode source volume(s) "snap create VOLNAME SNAPNAME", and perform an update with 7MTT. The success of the SVM-DR resync depends upon having an in-common 7-Mode snapshot.
2) Stop
the 7MTT service on the 7MTT box:
Via
DOS/PowerShell>
net stop "Transition Tool"
3) On
source cluster (clu1), we break the 7 to C SnapMirror, and remove volume
transition protect:
clu1::>
snapmirror break
-destination-path FAS101_clu1:TVOL001
set -c off; set d
volume transition-protect
-vserver FAS101_clu1 -volume TVOL001 -is-enabled off
4) On
the DR cluster (clu2), we create the SVM-DR vserver, vserver peer, and then
accept the vserver peer on clu1:
clu2::>
vserver create -vserver
FAS101_clu2 -subtype dp-destination
vserver peer create -vserver
FAS101_clu2 -peer-cluster clu1 -peer-vserver FAS101_clu1 -applications
snapmirror
clu1::>
vserver peer accept -vserver
FAS101_clu1 -peer-vserver FAS101_clu2
5) We
create the SVM-DR relation, initialize, and wait for initialization to
complete:
clu2::>
snapmirror create
-source-path FAS101_clu1: -destination-path FAS101_clu2: -identity-preserve true
snapmirror initialize
-destination-path FAS101_clu2:
snapmirror show
-destination-path FAS101_clu2:
6) Once
the SVM-DR relation is initialized, we break and delete it:
clu2::>
snapmirror break
-destination-path FAS101_clu2:
snapmirror delete
-destination-path FAS101_clu2:
7) Then we
release the SVM-DR relation on the source cluster, and resync the 7 to C
SnapMirror:
clu1::>
snapmirror release
-destination-path FAS101_clu2:TVOL001
snapmirror resync
-destination-path FAS101_clu1:TVOL001
Note: You have to do the delete and release,
otherwise the 7 to C resync errors with "Destination ... cannot be the
source of another SnapMirror relationship"
8) Restart
the 7MTT service on the 7MTT box:
Via
DOS/PowerShell>
net start "Transition Tool"
And add
in the credentials to the 7-Mode Controllers and ONTAP systems.
9) {When
the time comes...*} Complete the Transition via 7MTT!
Image: Completed 7 to C Transition
*You don’t want this to be too long after you’ve
seeded SVM-DR since the in-common Snapshots may be lost.
10) Use
sync-cron.ps1 from the following link to ensure cron schedules are in sync from
source to DR clusters:
11)
Re-create the SVM-DR relation and resync:
clu2::>
snapmirror create
-source-path FAS101_clu1: -destination-path FAS101_clu2: -identity-preserve
true
snapmirror resync
-destination-path FAS101_clu2:
IMPORTANT: If you are considering using this method, please test thoroughly in your environment first.
APPENDIX: Event wafl.snaprestore.revert
Of
course, for the SVM-DR resync to work, there’s got to be in-common snapshots
(otherwise it’s a full baseline resync again.) You might see events like the
below when the volume SnapRestore reverts to the common (previous) Snapshot
copy:
clu2::> event log show
-event wafl.snaprestore.revert* -instance
Node: clu2n1
Sequence#: 11211
Time: 9/6/2016 00:16:02
Severity: INFORMATIONAL
Source: config_async_0
Message Name:
wafl.snaprestore.revert
Event:
wafl.snaprestore.revert: Reverting volume
TVOL001@vserver:8d16297e-7364-11e6-b6c8-000c29a5823e to a previous Snapshot
copy with snapid 1.
Corrective Action: (NONE)
Description: This message
occurs during volume initialization when the system reverts to a previous
Snapshot(TM) copy.
I am a huge fan of yours.. I have done the transition with out of box methodologies.. Whenever I needed help, your blog was there
ReplyDeleteHi Srinivas, thank you for the comment. Cheers, VC
Delete