Unsolved
This post is more than 5 years old
6 Posts
0
1238
CG in disabled state
We are in the middle of reclaiming a fair amount of space on our VNX. In order to do so we are creating new LUNs and migrating the existing VMs onto those new datastores. Once the VMs have migrated successfully, we are then requesting our storage team to attach the RPA Host in Unisphere to the new LUN.Please let me know if journal volumes will need to be created in unison with the new datastore creation ?
I believe,we will continue to see these ‘disabled’ CGs during the course of this reclamation process?
Thanks
forshr
1.1K Posts
1
November 29th, 2016 13:00
Have you removed the journals from he CGs to create the disabled state?
dragonforce04
6 Posts
0
November 29th, 2016 19:00
No Journals are not removed
forshr
1.1K Posts
1
November 30th, 2016 00:00
I'm quite sure why the CGs are in a disabled state (possibly related to LUN removal from the RP SG) but what you could have done is added the new datastore LUNs into the existing CGs and then removed the old datastore LUNs post migration. At RP version 5.0 this prevents journal loss. However, this approach would mean that the RPAs would have captured and replicated the writes during the SvMotion process.
Whatever the case you now need to ensure that the old datastore LUNs are removed from the CGs and the new datastore LUNs are added. The existing journal LUNs already assigned to the CGs can be reused of course. You could of course remove the existing CGs with the old datastores and recreate new CGs with the new datastores.
Regards,
Rich
udaykr
4 Posts
0
November 30th, 2016 08:00
I agree with that. How about in one of the CG, you have allocated 10GB Journal volume on datastore-1 and later you allocated another 20GB journal log on datastore-2. After 1 hr, if we decommission the datastore-1, what will happen. Does the sync re-initiates again from beginning with comparing the data on both sides?
dragonforce04
6 Posts
0
November 30th, 2016 09:00
Thanks Rich for clarification . I 've another concern that I need little help with , Is there a way I could monitor proactively the CG states in the recoverpoint, a way I can be notified if there is a error in CG without loggin in everyday and checking the status . Kindly suggest ..
Thanks.
forshr
1.1K Posts
0
November 30th, 2016 12:00
Yes, you can't remove a journal volume/LUN from a CG without affecting the filesystem of the actual journal. This is determined as a disaster by the RP system and a full init will be needed after disabling and re-enabling the CG.
forshr
1.1K Posts
0
November 30th, 2016 12:00
Absolutely, you can set up alerts via SMTP to your e-mail address.