Alternating chains on extents with block cloning / expand SOBR with block cloning

[pirx]

Hi,

this is a bit tricky due to block cloning. I have SOBR's with 2-4 extents on different Linux xfs repo server which have enough free space to keep new chains.

1. we started with a repo with one extent and pointed all jobs to this repo, now we 'll get additional servers wich will expand the SOBR with new extents. What is the best way to do this? As block cloning only works on the same filesystem an active full will be necessary. I guess moving/dehydrating + copying the data for large jobs with 10-20 TB and 10 weekly GFS points will not be possible. So putting the existing extend in maintenance mode / seal it and starting a new active full job by job is the way to go?

2. Although we have copy jobs too, for redundancy reason it would be nice to have backup chains alternating on different extents (maybe once a month). Even If I would accept the fact that this results in a new active full, I don't see a way to configure it.

[Gostev]

Hi

1. Yes, sealing is the way to go. I believe the job will do an active full automatically in this case.

2. You're the first to request this in all the years SOBR existed so I doubt there will be enough demand for us to justify implementing this logic. By the way, how will up to one month old backups from another extent be useful for recovery? Also, if you're fine with double disk space usage from active fulls, then why not just do it properly: Backup Copy to another SOBR?

Thanks!

[pirx]

Yeah, it was probably a strange idea. It was mainly to reduce the impact when losing a single server. But with existing BCJ this is too complicated and too much overhead.

[Gostev]

Well, BCJ in the immediate copy mode are pretty straightforward... even if of course they can't beat the Capacity Tier Copy policy in terms of overhead