I have a backup job with 7 days retention policy but the job has accumulated 48 days. I know it should be for much 14 or 21 days for much, but 48 days seems insane.
How to figure out why the retention policy is not applying properly?
JosueM wrote:The restore points are all increment, except the firts one wich is full.
That's why restore points are not deleted - if you delete the first full, all the subsequent increments will become useless. You need to find out why periodic fulls are not created. For example, if the job is chained with other jobs and actually starts on the next day (not the one synthetic full is scheduled to be performed at), due to the time required to run earlier jobs.
I must confess I use more replicas than backups. But I have at least 3 backups now and all were created using the defaults.
No you point something that takes my attention, the job schedule and the full backup relation as I understand. This jobs runs daily at 23:50 , synthetic full is set to run each Saturday weekly.
Then it seems like the synthetic full does not perform.
Should I change something at the schedule? or make the job to run the active full instead synthetic ?
It does not matter how you create full backup either reading the entire VM data or creating a synthetic restore point, so I'm not sure that switching between active full and synthetic will help. Do you have any backup window defined by any chance? Do you have any backup job sessions triggered on Saturdays at all?
JosueM wrote:if so, then yes it did run ok many saturdays until the repository got full.
Well, this could be the reason of why active full cannot be performed. However, you should see the corresponding message in the job session log on that day and the job should actually fail.
ok, so for what I understand from your replys is: for backing up Veeam needs a repository size that is capable to handle: 1) initiall Full, plus 2) subsequent incremental for the days defined in the retention policy, plus 3) synthentic or active full.
So for this especifica case the VM is about 800GB size. Initial took about 400Gb, then incrementals where for about 7 Gb average per seven days, and addind the next synthetic or active full for 400gb. Plus a 100GB margin for data grow.
I will need about 1TB of repository to backup this machine considering 7 days retention policy and only one job per day. it sounds ok or I'm totally lost?
Your calculations are correct, however, additionally you would need space for 6 increments more, since previous backup chain (initial full + 6 subsequent increments) can be removed only when the new chain reaches the specified retention (7).
Just wanted to add that if you want to monitor free space on your repository and predict how much space you will need, take a look at capacity planning for backup repositories report which is available in Veeam ONE.
Well in that case I see no advantage to use backup over replica, instead I would make them all as replica cuz is more storage efficient. In the example of the 400Gb virtual machine, a replicate would take only the 400Gb plus about 100gb or less to the 7 restore points of the week, so veeam will use about 500GB in hd space using replication en about 1TB+ in hd space using backup for the same period of time.
As I told at the initial of the post that most of the critical virtual machines are replicated and I was wrong thinking that using backups instead of replica would become a more efficient use of storage.
That's not actually correct. Replica is stored in VMware native format, while backups are compressed and deduplicated and typically full backup requires less than 50% of the VM size (as you can see, in your case full takes about 400GB for the 800GB VM).
Space considerations aside, backup and replication serve different goals and cannot be opposed one to another. Besides, it is always a good practice to have more than one backup.
You are right with the numbers, I did mistake cuz the VM size is 800Gb not 400Gb as I wrote before. So in storage usage for this one week example both backup and replication are almost the same.
Will have a look to the options mentioned for Vitaliy to see if making backups does worth over making the replica.
I did a test backup and things seems pretty weird here. I set restore points to 3 and the synthetic full run any day, but after runing a few times the job veeam is not making the synthetic full and neither area any errors on the job log.
Full backup is performed only once per day (unless active full is triggered using the corresponding command on the job's shortcut menu). If the job has already performed full backup during that day, all subsequent runs will be incremental. As I can see on the screenshot, you trigger the job several times during one day.
So, full backup are not determined by the restore points only and that's new for me. I wonder if the documentation has this information, cuz been hitting my head against the wall cuz of this.
As Alexander has mentioned, being executed automatically (via schedule), full backup will be performed no more than once a day. By the way, may I ask you what is the use case of running synthetic full on daily basis or even more frequently? Thanks.
The above test is not for a use case, but I guess there should some some applicable scenarios, I did the test because I did believe Veeam was not respecting the retenttion policy.
But now I know that retention policy consider days + restore points. IMHO this behavior is a little bit confusing.
