#8704 Need to backup/copy data from equallogics to storinator
Opened 7 months ago by smooge. Modified 6 months ago

Describe what you would like us to do:

The equallogics contains a lot of data for COPR which would be useful to be on the storinator which will be moved to RDU2.


When do you need this to be done by? (YYYY/MM/DD) 2020-03-31



@praiskup @msuchy does it make more sense to get old data off the equallogics or sync down the new stuff from Amazon?

Metadata Update from @smooge:
- Issue priority set to: Waiting on Assignee (was: Needs Review)

7 months ago

I'd say that ATM we should setup rsync from AWS.

I mean, equallogics content is getting older and older ... and syncing that doesn't
make much sense as we basically confirmed that all data were successfully
transferred everything to aws.

We though need other form of backup than the AWS snapshots (which are done
on attached, and live devices, thus not really ideal). That should be cron doing
rsync to storinator.

So, we need a machine/instance to mount those storinator volumes to sync to first right?

Also, (and I can open seperate tickets on this, but might as well ask them quickly here):

  • Right now we are only monitoring that ssh on copr aws instances is working. Should we open nrpe and add that? Or just leave it at ssh for now?

  • Right now we aren't backing up any of the other copr machines config/etc. Do you just want to do that to the storinator as well and have it all in one place?

  • alternately, we could also look at just backing up to a S3 bucket? (which should be fast and still inside amazon) ?

I missed the update @kevin, sorry.

So, we need a machine/instance to mount those storinator volumes to sync to first right?

Yes.

Right now we are only monitoring that ssh on copr aws instances is working. Should we open nrpe and add that? Or just leave it at ssh for now?

The plan is to move the infrastructure from AWS soon. So for now ssh is enough.

Right now we aren't backing up any of the other copr machines config/etc. Do you just want to do that to the storinator as well and have it all in one place?

Yes. We need to have UIDs properly allocated for BE/FE/dist-git/keygen. But
we primarily need to resolve BE backups.

alternately, we could also look at just backing up to a S3 bucket? (which should be fast and still inside amazon) ?

We are doing manual live volume snapshots. That is done to s3 afaik. Do we
have some standard way of doing this on filesystem level?

May we help with this somehow?

So we will be moving the hardware to RDU-CC next week. I will be finishing up the builds and changes at that time.

We keep receiving questions about ppc64le in copr.., are (some of?) the ppc64*
hosts from former OpenStack part of the HW movement?

Yes. They will be shipped next week and hopefully we can bring them back up in the weeks they arrive...

Login to comment on this ticket.

Metadata