Agreed, we should switch to instance of to remove usage of == and to improve
code reliability.
[ Full content available at: https://github.com/apache/geode/pull/2918 ]
This message was relayed via gitbox.apache.org for
notifications@geode.apache.org
The reason we elected to use a new class scoped data source class was so that
in the future if which data sources are valid changes this test will still be
comparing against an 'invalid' data source class. We originally tried to
compare it against a Managed or XAPooled data source class but ran
We looked into changing the OK status from servers into ERROR to make it more
clear when individual servers have issues, but this causes problems with the
CliFunctionResult resolution in gfsh which disrupts behavior with the
"ifExists" flag. Basically, if all the servers return the ERROR status
After looking over this, there is no code path which would allow us to attempt
updating cluster config in the event the binding is not found within cluster
config.
@jchen21 The reason you might see those two messages together is in the event
that there are no servers but cluster config is
