New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Raft election using consul leave #9755
Comments
tested today with consul 1.8.8 same behavior.
this is the logs from the leader:
in the specific:
If the cluster has a leader why adding back a node triggers a re-election, it shouldn’t for the sake of the stability of the raft consensus. |
Hi @ltagliamonte-dd ! I just tried to repro this in a 3-server node DC(built via Vagrant) using Consul v1.9.5. When I performed If you were to run this on the later-released version of Consul, do you to still observe this behavior? |
@ChipV223 thank you for looking into this, locally i can still see elections happening, those are the step i'm doing:
|
Thank you for the information @ltagliamonte-dd ! This seems very environment-specific, particularly with docker since I was not able to repro this using Vagrant VMs. I'll try again using docker containers and report back when I get the chance |
@ChipV223 any update on this issue? |
Hi @ltagliamonte ! I've not been able to look into this as I've been involved with other tasks. The plan is to try to run my repro again using docker containers sometime this week and will update when I have any news to share |
@ltagliamonte We're investigating a report of a similar issue and wanted to follow up here to confirm a few details:
|
Hello @mikemorris
|
Hi @ltagliamonte-dd, this issue was fixed by hashicorp/raft#476, added to Consul in #11375, and released in 1.10.4. Thanks for reporting this. I'm going to close this issue now. Feel free to re-open it if for some reason this does not resolve the issue. |
@blake i did test with the steps i posted using 1.10.4 and i can still repro the issue. |
Hey @ltagliamonte, |
Hello Hashicorp folks!!
I'm using consul v1.7.2, in my setup I have 5 consul-servers.
If i gracefully stop one of the consul servers (not the current leader) using
consul leave && systemctl stop consul
and then i start the process backs up a new leadership election starts.If i just stop the process with
systemctl stop consul
a new election isn't triggered.Is this the correct behavior?
It seems wrong to me trigger an election when a non leader server is gracefully stopped.
The text was updated successfully, but these errors were encountered: