Rebalance is stuck again


We’re using the cluster of about 35 nodes running using 2.5.1 community edition, split into 7 buckets of different sizes.

Last couple of days rebalance gets stuck at some percentage (sometimes the same %% on every node, sometimes at different positions) and does not progress (at least it shows no progress for some 10 hours).

I understand the version is quite old but maybe someone has an idea? Previously we noticed that a bucket being > 85% full can be a reason for such stopping, at least after we have that bucket more memory the next rebalance could complete. Unfortunately this does not help here.