virBeaver is a personal blog about virtualization and software-defined datacenter (SDDC) which is maintained by me, Tim Maier. virBeaver is derived from the numerous nicknames around beavers, which I had because of my birthplace, Biberach an der Riss in Germany.
Yesterday I had a scheduled update of his Veeam Backup & Replication installation with one of my customers. We planned to go from version 9.5 Update 4b (220.127.116.1166) to version 10 GA (10.0.0.4461).
As usual, I created an encrypted configuration backup before the update for safety reasons. How this works and why you should encrypt the configuration backup you can read here and here. I prefer to be a little more cautious at this point, before I have the trouble in hindsight. However, I did not need the configuration backup. The update went smoothly and without problems.
Since I carried out the update during the day, it was not possible, in agreement with the customer, to perform a complete backup run directly after the update. Therefore, I did a short functional test using the Quick Backup capabilities of Veeam Backup & Replication. There were no problems here either.
Heute rief der Kunde an und berichtete von fehlgeschlagenen Sicherungsaufträgen. So I looked into it:
This week I reworked the backup of a customer. The customer is using Veeam Backup & Replication to back up their VMware vSphere infrastructure. For whatever reason (the customer couldn’t give me the exact reasons), a single backup job was created for each VM.
Apart from the poor deduplication rate, these over 100 backup jobs were anything but easy to manage and monitor. We looked at different ways to group the backups and after a short time on the whiteboard we came to a first solution: VMware vSphere Tags. These tags exist since vSphere 5.1 and are the successors of the Custom Attributes. Within Veeam Backup & Replication, they can be used as selection criteria for the VMs to be backed up in a backup job.
First we had to create the appropriate category (Backup) and tags within vSphere. Then we were able to use the newly created tags as selection criteria within the backup job.
The customer operates some hosted VMs for external customers. The VMs to be backed up are in different domains with different credentials. For this reason, we faced the challenge of having to use different credentials for different VMs within a single job. This is generally no problem for Veeam Backup & Replication. But when working with container objects like tags (but also clusters, folders, datastores, etc.), the functionality is a bit tricky to find. If we look at the settings for Credentials under Guest Processing, we first see only the container object for which we can assign certain credentials.
But if we click the Add… button and switch to VMs and Tags again, we have the possibility to expand our selected tag and see all VMs to which our tag is currently assigned.
If we now select all VMs for which we would like to assign individual credentials and press Add, we will be able to store the correct credentials in the next step.
Even if it is not immediately recognizable at first glance, the assignment of individual credentials works not only with the dedicated selection of individual VMs, but also with the use of container objects such as tags. This enables a highly automated implementation of backup jobs even in complex environments with different credentials. New VMs only need to be given the appropriate tag and the backup administrator must receive the appropriate credentials if more than just backing up the entire VM is required.