vRanger Backup Issues

[AlanTLR]

Within our environment, our VM admin is using vRanger to backup the VMs (not his nor my choice--we do have our own VMs that we are able to back up using NetBackup Enterprise Server 7.5.0.5), but they want the vRanger image files backed up.  I am able to back up the image files successfully upon incremental backups, but with full backups, they fail with status 13 or status 24 after going over 21 hours.  I want to split the backups up, but the vRanger is on a Windows Server 2008R2 machine, which saves the images to the E: drive with part of the VM name in the field.  Most of the VMs have a standard prefix for the site, but there are a few that do not have the same:

e.g.

  sitehost1

  sitehost2

  sitehost3

  annoyingname

  someothername

With only a few of these, I would be able to do a simple globular expression, but there are about 80% with the site name and 20% without and over 100 VMs, which makes doing a globular expression for the backup selection (or exclude list) a chore.  Beyond that, if a new VM is spun up, I have no way of guaranteeing that the new VM's image is backed up.  I'm doing this per policy, but would it be better to set up User backup and have the vRanger launch a backup once it's completed the image?  Or is there a better way to select the VM images to back up?

Thanks for your time, thought, and contribution.

NBU Master: 7.5.0.5 on Solaris 10

[RamNagalla]

1) creating mulitple Directories in the E dirve  with respective to site name or application name.... or some other common paramenter and backup them up with superate streams would be in the one option.

2) creatating the user-schedule and let the VMs admins to trigger the backup once Vranger compleated the backup.. they can also also automate it by using the 3rd party schedulers..

[RiaanBadenhorst · Accepted solution]

Hi,

Is vRanger logging the file names it created during the backup? If it is and you can access that file you could write something that runs through the files and then creates backup jobs in batches. A policy would need to exist for it to run against.

Thanks

Riaan

[AlanTLR]

The first option would require the VM/vRanger admin to do some kind of intervention, which I don't think they would want to do at this stage.  The second one is feasible and might have some push-back on the admins, but I might be able to get them to do it if I can show the benefits.

[AlanTLR]

I was thinking something similar to your solution, like a pre- and post- script, but it felt kind of "kludgey" (not that it necessarily is).  It doesn't necessarily log the files, but I have asked that the drive be shared out, so I could also track changes using my own server, instead of installing scripts on theirs, generating the policies' client lists daily.  This also feels kludgey.