When the client asks to insert data, the DGRAPH service drops a node, and when the node is restarted, the client keeps reporting “dgraph UNKNOWN: Please retry again, server is not ready to accept requests” ?
Is it normal for the server to respond when I stop the client, wait a while, and then request again?Why is that?
Without logs, we can’t tell much. But probably it is doing a balancing shard, not sure.
You can simulate a scene like this, 100% reproducible.Can it be solved?
Sorry, there was a typo in my comment. Also, I don’t see an issue to be solved. Can you be more specific?
Every time you restart an instance, it has to pass through all necessary steps to make the cluster available. So you will have some warnings like “Please retry again, server is not ready to accept requests” for a moment.