Rolling upgrade fails

Couchbase Server
1

Hi,

We’ve started a rolling upgrade of a 30+ nodes Couchbase cluster from version 2.5.1 to 4.5.0. After we’ve added the first 4.5 node and tried the first rebalance it repeatedly fails for a certain bucket with the following:

Rebalance exited with reason {unexpected_exit,
{‘EXIT’,<0.10966.6>,
{wait_backfill_complete_failed_for,
[{‘ns_1@hz-cb35.local’,
{‘EXIT’,
{{got_exit,
{‘EXIT’,<0.10927.6>,
{{badmatch,{error,closed}},
[{ebucketmigrator_srv,
upstream_sender_loop,1,
[{file,“src/ebucketmigrator_srv.erl”},
{line,719}]}]}}},
{gen_server,call,
[<0.10922.6>,wait_backfill_complete,
infinity]}}}}]}}}

and

killing myself due to exit signal: {‘EXIT’,<0.10926.6>,
{{badmatch,{error,closed}},
[{ebucketmigrator_srv,
upstream_sender_loop,1,
[{file,“src/ebucketmigrator_srv.erl”},
{line,719}]}]}}

Is there any ideas what’s going on?

Thanks,
Kirill