Slow reconnection after `Reset reason: local reset`

Issues Dgraph
1

What I want to do

I want to test whether the local service connection and data are normal after the dgraph pod deployed in k8s is deleted

What I did

After manually deleting the alpha or zero master node, a query request is initiated, and the local old connection will report an error. Then I reconnected, but it was very slow

image
image1298×748 196 KB

Dgraph metadata

dgraph version： v20.11.2

image
image1518×210 71.8 KB

dgraph version
2

alpha-1 log:

image
image1590×1106 542 KB

3

I converted this into an issue.

The main issue is that the reconnection is slow?

4

Hmm, yes, because I see that the alpha log should complete the master node switch quickly, but I don’t understand why it takes so long to reconnect, and it takes about 10 seconds to 40 seconds irregularly.