i would like your valuable regarding how to make a restore from TimeFinder Snap on a VMAX20K with SE 7.4 and 5875 code.
Well in a customer we have implemented the TF/Snap on their Billing System Application i think is 10g Oracle Cluster and on their ERP application 10g Oracle as well but with Oracle RAC implemented.
Both Clusters are HPUX 11.31 with Veritas Volume manager 5.0 if i remember well for the Volume manager.
Snaps are executed every 8 hours and we have 8 pairs.
The first pair is executing at 00:00, at 08:00 and 16:00
The second pair is execuing at 01:00, at 09:00 and at 17:00
If it is necessary how i will execute the restore procedure from the Snaps to the Production host?
If i terminate the Snaps, then will i lost any changes kept on the snaps?
How can i disable "let's say" the procedure in order to use all the available Snaps?
I mean that if i want to restore from the pair of 00:00 hours but the customer also wants to restore from the 23:00 (one hour ago) because the data on the snap of 00:00 are not the valid, how can i disable the pair of the 23:00?
And then what is the procedure for the restore?
1. Terminate the Snaps? (Customer uses txt files and not Device Group for the Snaps)
2. Put the devices in a masking view
3. Mounted the snaps on the mounted host
4. Unmount the file systems on the Production host
5. Deport the disk groups on the production host
6. Start the restore procdeure using the txt file and the symsnap restore command.
Your help will be valuable and much appreciated.
Thanks in Advance
i am not sure i understand your question but you definitely don't want to terminate snaps as you will lose your point-in-time copy. If you restored from snapshot that was created at 00:00 and then say ..nah data is not not good, let me try to restore from 23:00. Then you will need to terminate the "restored" (symsnap terminate -restored) session not the original 00:00 session. Is that what you are asking ?
First of all Dynamox thank you very much.
So my first question is answered and i had the same feeling that i didn't have to terminate the session.
So SNAPS will be remained in ACTIVATE -consistent state.
Let me paraphrase again your answer regarding the restore.
If i restored from a snap created at 00:00 and the data are not good and i decide to restore from the 23:00 or the 01:00 let's say, i have to terminate the RESTORED session and not the original.
i believe that there is a command to check if i have a restored session.
Regarding the numbered steps, what is your comment?
Is that correct?
Because the restore regards Oracle do i have to do something else?
Do you have any steps to suggest to me?
Thanks in advance
First of all you need to create the necessary Storage Groups and Masking Views for the 8 Snaps regarding the mounted host. As dynamox said you do not terminate the snap sessions!!!
You stop the application on production host and you unmount the production volumes.
Now you create/change the SG of the Masking View for the "mount" host to utilize a specific snap.
You start a vxdisk enable command in order to discover the disks.
You import all the DGs.
You start the application to check the specific snap.
If the specific snap is approved for restore then you stop the application on the mounted host and deport all the DGs.
Then you start a snap restore session to the production volumes using the specific snap.
When it's finished you terminate the specific session.
You can check the snap sessions using symsnap -sid xxx query -multi.
The original snap sessions will have CopyOnWrite state, the restore copy session will display as RestInProg or Restored state.
Also don't forget to check the SAVE area just in case.
I stongly suggest to check the pdf SE 7.5 Timefinder CLI.pdf from the documentation set p.73.