[teiid-issues] [JBoss JIRA] (TEIIDSB-64) Research HA

[Steven Hawkins (Jira)]

    [ https://issues.jboss.org/browse/TEIIDSB-64?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13721414#comment-13721414 ] 

Steven Hawkins edited comment on TEIIDSB-64 at 4/11/19 10:58 PM:
-----------------------------------------------------------------

Also it looks like we'll want to set service.spec.sessionAffinityConfig.clientIP.timeoutSecond for pg/jdbc - the default is 10800 seconds or 3 hours.   The maximum is 86400 or 24 hours, so that introduces a hard limit on session life if using more than one replica.


was (Author: shawkins):
Also it looks like we'll want to set service.spec.sessionAffinityConfig.clientIP.timeoutSecond for pg/jdbc - the default is 10800 seconds or 3 hours.  

> Research HA
> -----------
>
>                 Key: TEIIDSB-64
>                 URL: https://issues.jboss.org/browse/TEIIDSB-64
>             Project: Teiid Spring Boot
>          Issue Type: Quality Risk
>            Reporter: Steven Hawkins
>            Assignee: Steven Hawkins
>            Priority: Major
>
> Teiid's pg, and jdbc assume session affinity.  Also OData pagination works best with session affinity.
> For HA, and horizontal scaling, it would be good to have more than 1 pod alive and receiving traffic.  It appears that there is a sessionAffinity setting at a service level https://kubernetes.io/docs/concepts/services-networking/service/ and some work that can be done to customize routes https://docs.openshift.com/container-platform/3.9/architecture/networking/routes.html#routes-sticky-sessions - but that doesn't appear to work for just tcp traffic.



--
This message was sent by Atlassian Jira
(v7.12.1#712002)