My assumption is that a server should auto recover after sufficient successful connection checks.
Method to reproduce. (version 1.3.5)
1, Shut down MySQL on one node and wait for the connection checks to mark the node as SHUNNED
2, Start MySQL again.
3, Monitor monitor.mysql_server_ping_log and monitor.mysql_server_connection_log - no more errors being reported
4, Expected result. DB nodes status to change from SHUNNED to ONLINE - but simply remains marked as shunned.
5, restart proxysql or manually bring ONLINE
Thank you for the report. This is actually a duplicate of #971 .
The server is automatically brought back online when traffic hits the proxy.
The fact that currently is reported as SHUNNED it just means that the last time it was used was SHUNNED.
As pointed in #971 it is possible to improve the status reporting
Thank you for the prompt reply. That could explain things, as this is an isolated test environment where there is no traffic going through the proxy. I shall retest with some dummy traffic and I shall read up issue #971
@barrypowellpulsant This is my exact experience and expectation as well. Other proxy systems automatically update status of backend servers so I was expecting the same here.
Most helpful comment
Thank you for the report. This is actually a duplicate of #971 .
The server is automatically brought back online when traffic hits the proxy.
The fact that currently is reported as SHUNNED it just means that the last time it was used was SHUNNED.
As pointed in #971 it is possible to improve the status reporting