Hi Team,
There are two questions about High Availability need your help.
I found that sometimes the HA state of confd will transfer form “slave” to “none”, or from “master” to “none” by confd itself. I meaned the state was not set by api functions from app.
These are my question:
1. why the state of confd will transfer automatically form “master” to “none” ?
2. Is there anyway to prevent the state transition caused by confd? In our solution, we expect the app to get control of the HA state of confd.
Not sure what is happening in your setup but if ConfD lost connection to the other instance in an HA mode it will transition to the mode none and become standalone.
This looks like it will requiring some debugging that you would need support involvement for.
This is definitely the expected beahaviour when a slave loses contact with the master - it will transition to “none” since it can’t be a slave without a master, and wait for instructions from the HAFW (e.g. that it should become master because the master has died). But a master should not transition to “none” of its own accord - perhaps the ConfD instance was restarted? In that case the HAFW needs to instruct it to be master again.
Yes, if you can’t figure out why the master became “none”.