While running a backup Job’s that contain a large number of VM’s you receive random failures. However when these failed VM’s are run outside the backup job and outside peak backup time, these VM’s backup without issue.
While running a backup job that contains a number of VM’s, this amount of traffic can possibly outperform the storage of your VMFS. Its possible vRanger Pro 4.0 DPP can fill a single 1 GB NIC on a repository, with just three to five simultaneous backups at once. The same goes here for requesting a high amount of data from the VMFS to push thru a single connection to its repository.
If after upgrading to 4.1 or later you’re still experiencing repository read or writes errors, it is suggested you create a working baseline that works best for your environment. This can be found in the vRanger DPP GUI under Tools>Options>MyJobs>Configuration.
We recommend you change your simultaneous backup count to the following:
Per LUN = 3
Per Host =1
Per repository = 3
Once you have consistent successful backups based on what your environmental threshold can handle, you can start to raise the per-repository by increments of 1. If at that point you continue to show success you can increment by 1 the host count.
Remember to keep in mind that if you start to receive failed backups you will need to dial down these settings or add more Network or IO resources into your repository to be able to achieve the desired simultaneous backup count.