Closed eloo closed 6 years ago
Hi @eloo Currently the standalone ist only single instance, so loadbalancing will not work, as you have 2 indioendent instances. This could be solved using solr cloud with 2 or more servers and the configuration you can find in the repo. We will come back with a more detailed instruction!
Hi @eloo,
I've added a basic documentation on how to run chatpal-search-standalone in cloud mode based on Docker. I hope you can transfer this into a kubernetes setup - if you have any further questions please don't hesitate to ask.
pretty cool.. with test it asap ;)
@eloo Did you get the SolrCloud-K8S-setup up and running? We're looking into this as well. And it would be very useful, to find a no-memory-leak-setup for Chatpal on Solr in K8S - be it SolrCloud or a Master-Slave-Setup!
@mrsimpson sorry but no.. didn't have time for that atm
@eloo Do you have a setup with a running Solr-cloud already onto which you want to deploy Chatpal to or are you just looking for a failsafe way of setting up Chatpal on a k8s-cluster?
@mrsimpson i'm trying to create a HA setup for RocketChat together with ChatPal, so at least a i need two instances of ChatPal on the k8s-cluster yes.
Hello, i've tested the chatpal-search together with rocketchat and noticed that when i'm trying to run two instances of chatpal behind a loadbalancer (kubernetes setup) that some search queries will return an emtpy result but some of them are returning a valid result (keyword found). So i guess we need to configure the chatpal-search somehow? I've found for solr a zookeeper argument, maybe that could be a solution?
Do you have any hints how a multiinstance chatpal-search setup could look like? Thanks