Skip to:
When the pbm-agent (during the pitr procedure) fails on reading the storage config it may get locked out of starting the slicer (which results in missing pitr data for a given replicaset in sharded cluster). See https://github.com/percona/percona-backup-mongodb/blob/v2.5.0/cmd/pbm-agent/restore.go#L214-L226
It seems like pbm-agent for each replicaset member stops at https://github.com/percona/percona-backup-mongodb/blob/v2.5.0/cmd/pbm-agent/restore.go#L183-L185
I was able to get rid of the issue by releasing the lock before returning the error.
PR: https://github.com/percona/percona-backup-mongodb/pull/957
When the pbm-agent (during the pitr procedure) fails on reading the storage config it may get locked out of starting the slicer (which results in missing pitr data for a given replicaset in sharded cluster). See https://github.com/percona/percona-backup-mongodb/blob/v2.5.0/cmd/pbm-agent/restore.go#L214-L226
It seems like pbm-agent for each replicaset member stops at https://github.com/percona/percona-backup-mongodb/blob/v2.5.0/cmd/pbm-agent/restore.go#L183-L185
I was able to get rid of the issue by releasing the lock before returning the error.