Duplicate nodes in admin UI


I deployed docker images of cockroach and all nodes are connected. But I see duplicate nodes in the admin UI.
For some reason of network issue/deployment problem, 3 of the nodes got disconnected and admin UI displayed these as dead.
When network is restored, these 3 nodes are restarted. Then these nodes got added and admin UI displayed that these are healthy. But the problem is those same nodes which are restarted are shown twice in admin UI, once as dead and other as healthy with exactly identical IP:PORT (only node number varies). So let’s say, instead of 10 nodes admin UI displays total nodes as 13 and shows 3 as dead. But those 3 dead ones became healthy after restarting.

How to correct this? is this an expected behavior? Its really troubling to see dashboard with 3 nodes as dead always. Not a correct way of representation, please help.

Version used is v20.2.5

Hi @Messi, thanks for sharing a detailed description of your issue. I agree this is a confusing user experience. We do currently have some known issues with the node liveness statuses and we will certainly take this use case into account when working on this in the near future. There are a number of related issues to yours that we’ll be looking at in our next release.

I collected your description of the issue in this bug report: ui: node reconnection events confused with "dead" node events · Issue #61562 · cockroachdb/cockroach · GitHub

If you could add more reproduction steps (especially w/ docker-compose if possible) and a screenshot of the issue (the confusing node names showing as dead and live simultaneously) this would help us resolve the issue faster. You can also follow the bug in GitHub to be notified when we have a fix.