Details
-
Task
-
Resolution: Fixed
-
Major
-
7.2, Trinity
Description
What is the problem?
We recently had a customer issue where they were running three backup service nodes to do a cloud backup, each with a staging directory on a local drive. Because of the large number of previous backups, populating the staging directory took a long time which causing task overlap. Multiple tasks can run at once (MB-60676) so the user got a confusing error message about lockfiles.
What is the solution?
Until MB-60676 is fixed we should guide customers towards using one backup node. I don't think we should be too pushy here as multiple backup nodes are legitimate for resiliency reasons and if you want to take multiple backups at once (e.g. you have a repo per bucket) but in general customers should consider carefully whether it applies in their case before using multiple backup service nodes.