caching - Ignite in Kubernetes has problems finding servers -
we deploying four-instance ignite cluster in kubernetes along several micro-services rely upon it. servers host several caches, of read-through and/or write-behind. kubernetes service , clients configured discovery directed in ignite documentation.
often times on server startup @ least 1 instance never connect rest of cluster, , instance may need deleted , replaced multiple times before joins.
server instances may end segmenting out, 2 @ time, cluster no recognizable reason. kubernetes replace nodes results similar happens on cluster startup.
clients unable find server instance , remain unable connect cluster until entire server cluster restarted.
the cluster may run days w/o issue, server instances can't see each other, clients can't find servers, , spend day bouncing on , on until works. have tried several changes kubernetes service, client , server configs, unable diagnose, less repair issue.
we suspect specific kubernetes since similar configuration running fine in vm-based environment w/o these issues.
help appreciated.
Comments
Post a Comment