Help me understand sync jobs

[oester]

I have a single PBS instance with a 4TB local nvme datastore called "nvme1". What I want to do is sync (copy) that datastore to another directory (datastore) on that same PBS server that's mounted as "pbsbackup". This PBS server has both datastores (nvme1 and pbsbackup) defined

- I configured a "remote" pbs server that's actually the same PBS server under a different name
- I then created a sync job (push) that uses "nvme1" as the local datastore and "pbsbackup"

After the sync job runs it looks like the "remote (pbsbackup)" datastore has backups that have been removed from the "local (nvme1)" datastore.

I guess I'm not understanding what's the correct configuration of the sync job that I want.

What I want is to have all the backups that are in nvme1 replicated to pbsbackup, with data not in nvme1 removed from pbsbackup. Can someone help me get the correct configuration?

 

[Chris]

Hi,

What I want is to have all the backups that are in nvme1 replicated to pbsbackup, with data not in nvme1 removed from pbsbackup. Can someone help me get the correct configuration?

you can use the remove vanished flag to control that behavior.

 

[oester]

Hi,

you can use the remove vanished flag to control that behavior.

Yep, I understand that. Looking at the documentation a bit further I see this as the first line under the "Managing remotes and sync" section:

Sync jobs are configured to pull the contents of a datastore on a Remote to a local datastore.

That being the case, I think what I really to do is define the datastore "nvme1"as my remote datastore and "pbsbackup" as the local datastore and then configure the sync job on the "pbsbackup" datastore to sync to "nvme1" datastore.

 

[oester]

Changing the config as stated seems to have resolved it.

 

[Chris]

That being the case, I think what I really to do is define the datastore "nvme1"as my remote datastore and "pbsbackup" as the local datastore and then configure the sync job on the "pbsbackup" datastore to sync to "nvme1" datastore.

That however depends on what the source instance and what the target instance should be and in which direction you are performing the sync. You can of course set up a pull sync job to fetch the data from the remote (your destination target) instead of setting up a push on the source. Both variants are possible.

However, for both directions you can control if backup snapshots/groups/namespaces vanished from the sync source should be removed from the sync target by enabling/disabling the remove vanished flag in the sync job configuration.