we have a production Nexus Server with around 2TB of Space and two blob stores "default " and “Daily”
in the default blob stores, we have various repositories data and in the daily blob, we are storing artifacts from nighly jobs which are not needed to be backed up, for backing up the DB and Blobs we are using a scheduled task which uses a groovy script and takes the backup of DB and Blob “default” only, now when we are trying to restore this backup we have issues until we built the component DB index again, for testing we have created a test server with the same version of Nexus and restored the data, the restore works i.e. we can see all the repos but when the server restarts the blob default is not accessible but after doing a reindex it works somehow, i know kind of reason as we are taking backup of configuration for “daily” blob also but while restoring only restoring the corresponding default blob we might have a mismatch, if yes this is the reason, is there a way to only take the backup of configuration which contains the data for default blob or what do you suggest?
The backup tasks backs up everything, so my personal opinion is you’re overcomplicating the scenario by trying to split it.
My recommendation would be to split the servers instead if this works for your workflow, then you can just full back up one place without having to worry about hijinx and not backup the other at all.
As far as the reindexing, I’m not sure but I wonder if the backups (db & blobstore) are not taken at the same time. Reindexing (or repair) might be needed if there were discrepancies…it shouldn’t normally be needed as far as I know.
Ref https://help.sonatype.com/display/NXRM3/Prepare+a+Backup “The blob stores and metadata databases must be backed up together or the blob store metadata may be non-descript when attempting to restore the exported data.”
Some general thoughts,