Re: ClientProtocolService failures in geode-core dunit

2017-10-12 Thread Kirk Lund
I'd recommend "geode-api" (keep the module names and jar names all lower case). On Thu, Oct 12, 2017 at 3:12 PM, Galen O'Sullivan wrote: > Yeah, it probably shouldn't be logging the whole stacktrace. I think we > should log a message at debug level, and log a warning if a new client > tries to c

Re: ClientProtocolService failures in geode-core dunit

2017-10-12 Thread Udo Kohlmeyer
Currently there is a PR awaiting review, that will address this issue. https://github.com/apache/geode/pull/920 On Thu, Oct 12, 2017 at 3:12 PM, Galen O'Sullivan wrote: > Yeah, it probably shouldn't be logging the whole stacktrace. I think we > should log a message at debug level, and log a warn

Re: ClientProtocolService failures in geode-core dunit

2017-10-12 Thread Galen O'Sullivan
Yeah, it probably shouldn't be logging the whole stacktrace. I think we should log a message at debug level, and log a warning if a new client tries to connect when the new client protocol isn't available. If we made a "geode-public-API" module, we could avoid the cyclic dependency and do away wit