Hi We had an issue where one client joining consul agents from different projects to central project where they kept all servers. All agents using local service account but using end points approach to connect to remote consul server. Remote consul service has ingress IP attached.
Flow: Project1 --> Local service account --> end point/Ingress IP of remote server --> Consul server (pet set) Using above approach, consul agents throwing unable to connect sometimes (not always) and behavior is inconsistent. If we remove the local service account and directly use external ingress IP (another project in same cluster), join always successful. We did a below change and increased ARP table size fixed the change. Want to confirm whether this has any impact to cluster network in future or any side affect? https://github.com/hashicorp/serf/issues/263 https://trello.com/c/DZb8ghlZ/228-5-scale-document-tuning-options-for-arp-cache https://www.serveradminblog.com/2011/02/neighbour-table-overflow-sysctl-conf-tunning/ -- Srinivas Kotaru
_______________________________________________ dev mailing list [email protected] http://lists.openshift.redhat.com/openshiftmm/listinfo/dev
