To put this in context: we are migrating from Nexus 2 – OSS 2.14.18-01 to Nexus 3 – OSS 3.23.0-03.
We had to cancel the first and second migration attempts and we wonder if these incoherent blobstores’ values could be related to those cancelled migrations. Also, we have noticed the migration is going much slower at each attempt.
Thank you Matthew! You are totally right, we misunderstood the parameters.
But we still see something weird: In one repository “Total size” is 2726 GB but if we execute “du” command in the filesystem we see the repository only ocuppies 800GB, how this can be possible?
How I said, we are migrating from 2 to 3, and we had to cancel the first and second attempts. The migration is going slower each time. We would like to know if the generated nexus3 metadata of those canceled migratons could remain in the filesystem (.bytes and .properties) affecting the performance of the current migration execution due to the big number of files.
We have 100 millions of files in one repository (68 millions in the content folder where we see the metadata).
When we have canceled a migration we have removed this two folders as the doc says:
Shut down Nexus Repository Manager 2 and remove the entire NXRM 2 [ $work-dir/db/migrationagent ] directory.
Shut down Nexus Repository Manager 3 and remove the entire NXRM 3 **[$data-dir] directory.
But we wonder if we also would need to delete the metadata before starting a new migration attempt.
Note: we recently opened other thread about this slowness issue we have: