Is there a way to manually trigger a redistribution of files/pins across ipfs-cluster nodes? I’ve worked with Ceph clusters before and I think it has an auto-balancing mechanism when an OSD node goes down or becomes unhealthy. What is the equivalent of this for IPFS clusters. I do know that we can set the disable-repinning config variable to
false in the service.json config. But seems as though it only takes place for certain situations.
What if something is only pinned to one node, and THAT node is the one to go down? Will re-pinning commence? There would have to be blocks of that same data on other cluster nodes even if they were not pinned previously. Seems like it could be a precarious position to be in if some of our data only has a replication value of 1 (one node).
Ultimately, I’d like a way to be able to manually trigger a balancing of data before a catastrophic situation. So far I think the closest thing to this operation would be to manually go through the data and pin things manually to other nodes.