Backup Copy - Archival Points vs Many Incrementals

[Crof]

My goal with my off-site backups is to use them primarily as a 2nd backup, but also as a place for storing 'longer term' backups. Using the archive option is using a lot of space with full backups so I was wondering what others have done. I'm thinking about just increasing the amount of stored incrementals to cover what my archives would have.

Originally I setup my backup copy jobs to run daily for 2 weeks, and then checked "Keep the following restore points for archival purposes" and set it up for weekly backups for 4 weeks, Monthly for 3 months and 2 quarterly. I was aiming to keep about 6 months of backups off-site.

I expected it to do a base full, with incrementals all the way back - but it appears to have a base full with the 2 weeks of incrementals, and then a new full per archived backup - my math may be off, but in my case that would eventually be 7 full backups per copy job just for archives. Taking my one server with 2TB of used space for example, that would be 14TB of [undeduplicated] space used for backups.

In my case, the change rate on that data is VERY small - around 1.5% per week. So I'm thinking of instead just changing the retention to be 180 daily incrementals for a total of around 1-2TB. Is there a reason to not do this?

[PTide]

Hi,

So I'm thinking of instead just changing the retention to be 180 daily incrementals for a total of around 1-2TB. Is there a reason to not do this?

The longer your chain is - the higher are odds that it gets corrupted. One corrupted incremental backup will render all subsequent restore points unavailable for restore.

Thanks

[foggy]

That said, provided you do SureBackup on the original jobs and health check is enabled on the copy job, you're good to go.

[Crof]

Hrm...backup health is something to take into account for sure.

I've been looking around other threads since posting this and they noted recovery performance may be an issue for long incremental chains. Honestly reverse incrementals would be my preferred solution, but I'm 99% sure that doesn't exist for backup copy jobs.

I'm thinking of perhaps 'splitting the difference' and doing two copy jobs to the same off-site repository. One job would do 30 daily's to cover my short term needs, the other would do 25 weeklies to cover longer term retention. I would have 2 off-site fulls, but not 7.

[PTide]

You're 100% right - there is no reverse mode for backup copy jobs.

Regarding the plan to split the copy job in two - that won't help a lot, the overall size will likely stay almost the same. You can check that with a restore point simulator, kudos go to Timothy. If you want to save some space then the best option will be to use a dedupe storage as an offsite repository.

Thanks.