We have an issue with writing operational data with confd version 7.7.9.
We have open session to confd and when we want to write something, in previous versions we could just write it.
With version 7.7.9 however, when we want to do exact same thing and it happens that a new secondary just came online and is syncing with primary, writes fail with error ‘locked’.
We suspect that this change has something to do with it:
- ha: A write action with lock from southbound API during a new slave lock
caused an internal error in CDB when the new slave was unlocked. As a
result, the CDB subsystem has been updated to not allow any write action
with lock during the new slave lock.
My question is, is there a way to work around this that would not require us to rewrite all daemons to check for this error? Something like 'use this command with parameter ‘wait’ and it will wait for ha sync to complete?