Hey folks,
quick question: I have a 3 node cluster (HC) in a production
environment. I recently swapped all the storage SSDs with bigger
enterprise grades SSDs. So I took a host down (migrating all VMs away,
maintenance, shutdown, fix raid, gluster repair, back online).
This worked well and all three nodes are now hardware-wise replaced. The
last one is still gluster-healing (roughly 20% done). The other two
nodes are happily serving the vms while the third one is still in
maintenance.
my question is: Can I online the maintenance node while gluster is still
healing? Or will will open up a can of worms I would not like? It would
be cool having all three nodes back online soon while the gluster is
still repairing.
We have no urge tho. If there is a risk I would wait. If this is
somewhat risk-free and do-able why not :)
--
with kind regards,
mit freundlichen Gruessen,
Christian Reiss