Hello! I have Longhorn running on a Harvester clus...
# longhorn-storagep
powerful-easter-15334
09/01/2025, 6:11 AMHello! I have Longhorn running on a Harvester cluster, without too much load. The LH version is 1.7.2.
I have around 10 degraded volumes which are missing a replica. So naturally, LH will start rebuilding them.
The new replica will be rebuilt for quite a bit, then suddenly stop and will turn gray as either "running" or "stopped".
Are there any settings I should tune so the rebuild process can actually complete?
powerful-easter-15334
09/01/2025, 6:15 AMProbably relevant - I'm running on a somewhat resource-constrained environment with HDDs and 1G networking.
powerful-easter-15334
09/01/2025, 6:16 AMI decreased Concurrent Replica Rebuild Per Node Limit to 1, hoping this will lead to less congestion but that's not helping (enough?)
powerful-easter-15334
09/01/2025, 6:47 AMOoooh I changed Timeout between Engine and Replica from 8 to 20 seconds and a replica did actually succeed to get built, this is cool.
powerful-easter-15334
09/01/2025, 7:35 AMWell that success was short-lived 😄
powerful-easter-15334
09/01/2025, 10:59 AMHmmm, it seems back to normal now. Weird. Either way, cool, all my volumes have been replicated
✅ 1
👍 1
2 Views