Hey everyone, From what I've understood, when a storage node goes down for a reason some of the objects might have just one copy instead of the default 2 copies.
I realize it's by design but if I know that this storage node for some reason is going to be down for too long (2-3 days), is there some way to "trick" the ILM or make it somehow replicate the objects that have just one copy to a new storage node?
To recover missing copies of objects when a node fails, we use foreground verification, but I do not recommend it for your particular situation.
When a storage node fails and cannot be recovered, we replace that node and run foreground verification to re-evaluate all data in the GRID to make sure it meets the active ILM policies.
As foreground verification runs, the process replicates any missing copies to online nodes in the GRID consuming space on those nodes. After the original node is back online, I am not sure if running the process again would delete extra copies of data because the process is designed for recovering data from a permanently lost node.
Please also note that foreground verification takes days or weeks to complete so I do not recommend starting the process when a single node is offline for a few days.