-
Type: Bug
-
Resolution: Fixed
-
Priority: Major - P3
-
Affects Version/s: 3.3.3
-
Component/s: None
-
None
-
(copied to CRM)
-
Empty show more show less
As shown in basic uses of the sdam_viz tool, the unified topology seems to have a serious bug where replset nodes which are lost are not rediscovered when they return to the cluster. This seems to be related to the monitoring loop being disabled since monitoring commands and normal commands both emit error events from the Pool. I also suspect this is highly related to users getting high degree cases of Server selection timed out messages, since selection will necessarily time out if no servers are ever available.
- related to
-
COMPASS-3728 Top in performance tab times out when connected to Atlas
- Closed
-
COMPASS-3849 Compass loses connection after a while
- Closed