You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@hbase.apache.org by "Andrew Purtell (JIRA)" <ji...@apache.org> on 2010/04/08 21:03:36 UTC
[jira] Created: (HBASE-2429) [EC2] document how to set up elastic
load balancing across Stargate instances
[EC2] document how to set up elastic load balancing across Stargate instances
-----------------------------------------------------------------------------
Key: HBASE-2429
URL: https://issues.apache.org/jira/browse/HBASE-2429
Project: Hadoop HBase
Issue Type: Sub-task
Reporter: Andrew Purtell
Priority: Minor
Previously ELB across Stargate instances was made difficult (impractical?) because scanner state is not distributed among Stargate instances. I thought about making scanner state serializable and have multiple stargate instances retrieve and sync scanner cursors up in Zookeeper but obviously that is problematic for several reasons:
- Performance
- ZK write loading
- As a basic matter of architecture scanners are not meant to be concurrent across multiple clients
- The prospect seemed a lot of work
Anyway, today AWS announced ELB now incorporates session affinity:
http://aws.typepad.com/aws/2010/04/new-elastic-load-balancing-feature-sticky-sessions.html
http://docs.amazonwebservices.com/ElasticLoadBalancing/latest/DeveloperGuide/index.html?US_EnableStickySessionsLBCookies.html
So unless there is a failure of the Stargate instance (invaliding the scanner, but the client can restart on that rare error) ELB will transparently use cookies to maintain a session between a client and the Stargate instance, enabling proper function of scanners without further work.
Document this.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Resolved: (HBASE-2429) [EC2] document how to set up elastic
load balancing across Stargate instances
Posted by "Andrew Purtell (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HBASE-2429?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Andrew Purtell resolved HBASE-2429.
-----------------------------------
Resolution: Not A Problem
Closed per HBASE-2543.
> [EC2] document how to set up elastic load balancing across Stargate instances
> -----------------------------------------------------------------------------
>
> Key: HBASE-2429
> URL: https://issues.apache.org/jira/browse/HBASE-2429
> Project: Hadoop HBase
> Issue Type: Sub-task
> Reporter: Andrew Purtell
> Priority: Minor
>
> Previously ELB across Stargate instances was made difficult (impractical?) because scanner state is not distributed among Stargate instances. I thought about making scanner state serializable and have multiple stargate instances retrieve and sync scanner cursors up in Zookeeper but obviously that is problematic for several reasons:
> - Performance
> - ZK write loading
> - As a basic matter of architecture scanners are not meant to be concurrent across multiple clients
> - The prospect seemed a lot of work
> Anyway, today AWS announced ELB now incorporates session affinity:
> http://aws.typepad.com/aws/2010/04/new-elastic-load-balancing-feature-sticky-sessions.html
> http://docs.amazonwebservices.com/ElasticLoadBalancing/latest/DeveloperGuide/index.html?US_EnableStickySessionsLBCookies.html
> So unless there is a failure of the Stargate instance (invaliding the scanner, but the client can restart on that rare error) ELB will transparently use cookies to maintain a session between a client and the Stargate instance, enabling proper function of scanners without further work.
> Document this.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.