wenningerk
wenningerk
Thanks for the update. Unfortunately github doesn't create an email-notification upon an amend-commit. That is why I just now stumbled over the update by manually polling. Thus it usually makes...
Aah one more question arises looking at that error-handling. If we just close down the cmap-tracking we wouldn't be updated about changes anymore but proceed otherwise which might be dangerous...
> @wenningerk I was thinking about this but currently all CMAP failures (see https://github.com/ClusterLabs/sbd/blob/master/src/sbd-cluster.c#L197) are just "warning level" and don't trigger reconnect. That isn't entirely true as up to now...
> @wenningerk OK, so shall we really do a full reconnect (i.e. `sbd_membership_destroy()`) or just try to connect to cmap again? I'm thinking about creating another timer when cmap connection...
> onnection is lost while CPG is still working fine?
Sorry that was the wrong button ...
> @wenningerk We experienced this CMAP connection loss a lot in our CI but couldn't identify the exact conditions why this was happening. I thing it might have been cause...
> @wenningerk as I understand, the cpg reconnect via restart is not as simple to implement as it seemed (for sure not with my level of expertise). How shall we...
> Let's first confirm the concept. I'll ask the user to test it as well. Definitely something that looks as if it need improvement. Personally I guess we should try...
@gao-yan what are you using to simulate the stall behavior? Maybe we can add it to the tests then.