-
Type: Bug
-
Resolution: Fixed
-
Priority: Major - P3
-
Affects Version/s: 4.4.5, 5.0.3
-
Component/s: None
-
Fully Compatible
-
ALL
-
Service Arch 2022-10-03, Service Arch 2022-10-17, Service Arch 2022-10-31, Service Arch 2022-11-14
-
(copied to CRM)
We have a 3-node replicaset where I'm trying to replace a node with a new server. To do this I'd like to add the new server to the replicaset and remove an old one eventually.
I'm trying to add a member to this replicaset, but I see a ton of errors in the syncing replica member's log.
"NotWritablePrimary: Not-primary error while processing 'find' operation on 'database_production' database via fire-and-forget command execution."
On the primary server, where the replica is syncing from I can see a ton of:
"Dropping unhealthy pooled connection","attr":{"hostAndPort":"main-mongo-ams1:27017"}}
The machines can connect to each other and it seems some data is being synced, albeit very slowly.
The primary runs 4.4.5 and the secondary 4.4.5, but I also tried 4.4.9 with the same results.
Any idea what's going on?
- is related to
-
SERVER-70510 Avoid considering recovering nodes as electable
- Backlog