We have VDR 1.2 up and running backing up a number of VM's. As a test I created a backup job with a custom retention time to keep the last 3 backups and 1 full weekly backup. Unfortunately the entire past week seems to have been kept. Is my interpretation of the way the retention policy works is wrong or is something else fundamentally wrong?
The reason for try this custom retention time out is that we would like to start putting development machines on a to custom "very few" retention policy.
A reclaim operation needs to run (it runs automatically once a week) in order for the retention policy to be applied. I think you have not gone through the cycle to start the reclaim operation...thus the increased number of restore points in the interim. Another way to force the destination disk to smaller, so that in order to fit the number restore points, older onces need to removed for new ones to be added.
My vote is to let it run a few weeks and see if the reclaim operation kicks in to apply the retention policy.
This means 5 backups for 2 weeks.
So WHEN the reclaim is done, AFTER 2 weeks has past, the 6th backup is deleted at that time.
You are saying that EVERY backup for this group is still there even after 2 weeks.. and each of them has at least 6 backups?
Unfortunately the entire past week seems to have been kept. Is my interpretation of the way the retention policy works is wrong or is something else fundamentally wrong?
I think you missed something. 1 FULL week. so the backups don't get deleted UNTIL there is something to replace it with. So the second full week, AFTER it completes, now you have 2 FULL weeks. Then AFTER reclaim is done, at that time the PREVIOUS week will be deleted, not after time but after a NEW backup replaces it.
Weekly backups should show 1, the rest should ALL be ZERO. 3 recent backups is kept is also set. If there is anything else in the configuration, it will not be as you requested.
But it keeps all backups
I know it's supposed to work, but what if you mark the older ones as deleted, does it automatically delete them after that?
Otherwise this looks like yet another bug in VDR, which from what I read was SUPPOSED to be fixed in 1.1.
The retention policy that was actually applied can be verified in the datarecovery.log.
1. Identify the date and time of the last successful reclaiming operation.
2. Locate the datarecovery.log that corresponds to that time period.
3. Open the log and search for the following:
The retention policy, in the format "n/n/n/n/n", will be displayed for each VM.
Beneath that line, the results of the policy will be displayed for each restore point for that VM. (E.g., "meets last backup policy", "meets weekly policy", "will be deleted", etc.)
In your case, a retention policy of "5/2/0/0/0" should be displayed in the log.
If that's not the case, something may be amiss.
Questions regarding your screen shot of the Restore tab:
When was the last reclaiming operation successfully completed in relation to the screen shot?
Was the backup job's original retention policy modified?
Were the VMs protected by more than one backup job?
Were the VMs protected by a backup job that was subsequently deleted?
Was the dedupe store imported from another VDR appliance?
A policy of 5 Most Recent, 2 Weekly, 0 Monthly, 0 Quarterly, 0 Yearly (5/2/0/0/0) should retain the following after a successful reclaiming operation:
The 5 most recent backups.
From the remaining restore points, the two that are closest to the ends of the last two weeks.
Note that VDR bases the end of the week as Friday at 8 PM.
Following a successful reclaiming operation and prior to any more backups, the Restore tab should display a total of 7 restore points.