cancel
Showing results for 
Search instead for 
Did you mean: 

Best practice to backup CACHE DATABASE residing on a VM

ROCKY1978
Level 3

ALL,

I have a requirement to STOP the CACHE DB ,take a snapshot using NBU VMWARE policy type and backup the data from snapshot and START the CACHE DB once the snapshot is taken.

Since i am using VMWARE policy type i cannot use the bpstart & bpend scripts to start and stop the CACHE DB as there is no NBU software on the client to place these scripts.

Master,Media and client are on NBU 7.7.3 version

Is there a way this could be achieved while i could still leverage the VMWARE policy type.

1 ACCEPTED SOLUTION

Accepted Solutions

Assuming we are talking about a vm hosted on vmware, you can use a pre freeze and post thaw script to control cache. 

More info on this can be found on:

https://kb.vmware.com/selfservice/microsites/search.do?language=en_US&cmd=displayKC&externalId=10066...

If this is the same type of caché database we are talking about, you can use the following commands to let caché know you will freeze and thaw:

cache -s. -U%%SYS ##Class(Backup.General).ExternalFreeze()

cache -s. -U%SYS ##Class(Backup.General).ExternalThaw()

View solution in original post

7 REPLIES 7

Genericus
Moderator
Moderator
   VIP   

We have CACHE on AIX, and we snap using our storage. Granted we only have  4TB and 6 TB of data, but it is spread over several file systems. Since we set up each data set as a seperate LUN, we can snap them and mount them on a different server, then back up from there.

We are moving to Linux VM, but I think will will be retaining this config, so we have ease of backup and restore.

Plan A for DR does not use Netbackup, we recover from storage/snap, but we do the backup to tape for retention purposes. And we have a shadow server that we repopulate using restore.

 

NetBackup 9.1.0.1 on Solaris 11, writing to Data Domain 9800 7.7.4.0
duplicating via SLP to LTO5 & LTO8 in SL8500 via ACSLS

Genericus
Moderator
Moderator
   VIP   

To clarify - we use a controling script outside of netbackup to stop and clone and restart CACHE, then mount the cloned/snap disk on another server, the data is then backed up from there using schedules.

NetBackup 9.1.0.1 on Solaris 11, writing to Data Domain 9800 7.7.4.0
duplicating via SLP to LTO5 & LTO8 in SL8500 via ACSLS

Assuming we are talking about a vm hosted on vmware, you can use a pre freeze and post thaw script to control cache. 

More info on this can be found on:

https://kb.vmware.com/selfservice/microsites/search.do?language=en_US&cmd=displayKC&externalId=10066...

If this is the same type of caché database we are talking about, you can use the following commands to let caché know you will freeze and thaw:

cache -s. -U%%SYS ##Class(Backup.General).ExternalFreeze()

cache -s. -U%SYS ##Class(Backup.General).ExternalThaw()

Thank you for the info this looks good to me 

Thank you for the info.In my case the customer wants Netbackup to stop and start the CACHE so i would go with the below option 

https://kb.vmware.com/selfservice/microsites/search.do?language=en_US&cmd=displayKC&externalId=10066...

 

 

That is exactly what we are trying to do with NetBackup.   How do you call the controlling script and then have NetBackup wait until the clone is mounted before running the backup?   Please be as detailed as possible.

Thank you in advance.

Genericus
Moderator
Moderator
   VIP   

Since we have production cache server, and we run a separate batch reporting server, and a separate Netbackup server, our process is complicated...

List of scripts we run - called by operations job control

  1. script to stop cache
  2. script to unmount batch server filesystems
  3. script to unmount backup server filesystems
  4. snap/clone production to backup and batch copies
  5. relink batch
  6. relink backup
  7. start cache
  8. mount batch
  9. mount backup
  10. call user directed backup of cache from backup server or wait for scheduled Netbackup job.

 

OPS monitors the jobs, to make sure each step completes, and then the backups - watch for 32bit backups, that indicates the filesystems did not mount up correctly...

 

 

 

NetBackup 9.1.0.1 on Solaris 11, writing to Data Domain 9800 7.7.4.0
duplicating via SLP to LTO5 & LTO8 in SL8500 via ACSLS