Best Backup Method for 30 day + 6 months?

[Vinsanity904]

Hello,

I am trying to come up with an optimal configuration that uses the LEAST amount of disks space for our backups.

We'd like all backups to be 30 days rolling over and have 6 months worth of weeks ends. (so 24 weekly's outside of the current month).

The idea is, I can restore for any day within the past 30 days, however outside of those 30 days, you can only go to the weeks end.

What configuration would you all recommend for this? Backups are going straight to disk, then replicated via disk to disk replication. No Tapes involved. We're currently using forward incrementals. Backups are M-F, with an active full every Saturday.

[Vitaliy S.]

Hello,

I would recommend this configuration:

1. backup jobs with forever incremental mode (neither synthetic, nor active full backups will be required on the weekend) and retention policy set to 30 restore points.
2. configure backup copy jobs with the required retention policy. For example, set "copy every" interval to 7 days with a retention policy configured to 24 restore points.

Thank you!

[Vinsanity904]

Thank you very much for this recommendation, this sounds like a good idea.

A few quick questions.

1. What if I have a job configured with say 90 retention points right now, running backups M-F and an active full on Sat. in forward incremental mode.

How difficult would it be / what would be involved in converting this job to the 30 retention points, with forever incremental, do I simply re-configure the existing job and it self corrects?

2. For the backup copy job, is this essentially a FULL backup, or is this simply a backup of ONLY data that has changed for 7 days?

Thanks a bunch!

[foggy]

1. What if I have a job configured with say 90 retention points right now, running backups M-F and an active full on Sat. in forward incremental mode.

How difficult would it be / what would be involved in converting this job to the 30 retention points, with forever incremental, do I simply re-configure the existing job and it self corrects?

You can just re-configure the job, retention will apply at the end of the next successful run, adjusting the number of restore points on disk to the new value.

2. For the backup copy job, is this essentially a FULL backup, or is this simply a backup of ONLY data that has changed for 7 days?

Backup copy job is always incremental (copies only changed data).

[veremin]

2. For the backup copy job, is this essentially a FULL backup, or is this simply a backup of ONLY data that has changed for 7 days?

But if you want those points to be stored as full backups, put GFS retention scheme into use. Thanks.

[Vinsanity904]

Hello,

I would recommend this configuration:

1. backup jobs with forever incremental mode (neither synthetic, nor active full backups will be required on the weekend) and retention policy set to 30 restore points.
2. configure backup copy jobs with the required retention policy. For example, set "copy every" interval to 7 days with a retention policy configured to 24 restore points.

Thank you!

Thanks everyone for the responses, I want to make sure I understand the backup copy job correctly as well.

There are 3 main configuration points

1. "Copy Every" - If I set this to 7 days, every 7 day's its essentially going to just say "okay what does the data look like on day 7 and copy it?" For example, if data has changed all during the week, when the backup copy job runs on the 7th day, does it capture all of these changes that happened during the week?

2. If I want 6 months (24 weekly's) I enable "Keep the following restore points for archival purposes" and change "Weekly" backup to 24?

3. "Restore points to keep" from what I am reading, once the 8th of 24 Weekly's are kept the chain is transformed. So technically after 24 weeks, I should expect to have 7 retention points on disk that contain 24 weekly's worth of backup data technically?

[veremin]

Does it capture all of these changes that happened during the week?

It captures the latest chain state.

If I want 6 months (24 weekly's) I enable "Keep the following restore points for archival purposes" and change "Weekly" backup to 24?

Nope, you don't have to enable that option. Set copy every interval to 7 days and "restore point to keep" to 24.

Thanks.

[Vinsanity904]

Thanks for the reply!

My concern with the new forever forward incremental is to me IMO Not running a full periodically adds to the complexity of things. Longer backup chains with no full have more variables when it comes to rolling them up into a full for a copy job, for example. I would want to double check with Veeam that we can create a weekly full copy with forever incremental selected on the primary job.

Or is this incorrect analysis?

[foggy]

What probably confuses you is that the backup method selected for the source regular backup job does not affect what is being copied by the backup copy job. Backup copy job will always copy all the blocks required to build the new restore point that will correspond to the latest VMs state (the same state the last restore point in the regular backup job chain contains).

Also, you can achieve weeklies with backup copy job in two different ways. In case you set the "Copy every" interval to 7 days and set its retention to 24 restore points, you will have one initial VBK (full restore point) and 23 subsequent increments. If you need each of the weeklies to be stored as a full backup, then you need to enable "Keep the following restore points for archival purposes" and set "Weekly" to 24.

[Vinsanity904]

What probably confuses you is that the backup method selected for the source regular backup job does not affect what is being copied by the backup copy job. Backup copy job will always copy all the blocks required to build the new restore point that will correspond to the latest VMs state (the same state the last restore point in the regular backup job chain contains).

Also, you can achieve weeklies with backup copy job in two different ways. In case you set the "Copy every" interval to 7 days and set its retention to 24 restore points, you will have one initial VBK (full restore point) and 23 subsequent increments. If you need each of the weeklies to be stored as a full backup, then you need to enable "Keep the following restore points for archival purposes" and set "Weekly" to 24.

Fantastic, thank you for the explanation. I understand the differences between the two backup copy job configurations. However, I'm still "foggy" on what the pro's and con's are between the two. Do I get the same granular restore capabilities for both options? (i.e. with Option 1, can I go to the 14th increment and restore data from any given day for that week or no?)

Option 1: Copy Every 7 days, 24 restore points. 1 Full + 23 increments.

Option 2: "Keep the following restore points for archival purposes" and set Weekly to 24 each will be stored as a FULL backup. (Why would someone need this?)

[foggy]

Do I get the same granular restore capabilities for both options? (i.e. with Option 1, can I go to the 14th increment and restore data from any given day for that week or no?)

Yes, restore capabilities are exactly the same.

Option 1: Copy Every 7 days, 24 restore points. 1 Full + 23 increments.
Option 2: "Keep the following restore points for archival purposes" and set Weekly to 24 each will be stored as a FULL backup. (Why would someone need this?)

General recommendations for backup copy are to have some number of daily restore points in a form of regular incremental chain (hence, default "Copy every" interval value is 1 day) and then GFS restore points for each required period (weekly, monthly, etc.) as separate independent VBK files. However, if you don't need daily points at your secondary location, you can adjust "Copy every" interval for weekly backups.

[Vinsanity904]

Thank you! Case closed! I appreciate everyones time on this!