[
https://issues.apache.org/jira/browse/GEODE-66?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15089653#comment-15089653
]
Swapnil Bawaskar edited comment on GEODE-66 at 1/8/16 6:32 PM:
---------------------------------------------------------------
Hmm, good question.
>From what I have seen and specially since more and more geode deployments are
>going to be in the cloud, the servers are/going to be configured identically.
>Which means that resource management is a cluster level setting as opposed to
>a server level setting.
I think we need to remove the eviction and critical heap settings from the
`start server` command and instead create a new gfsh command
`configure resource-manager --eviction-heap-percentage=80
--critical-heap-percentage=95`
was (Author: swapnil.bawaskar):
Hmm, good question.
>From what I have seen and specially since more and more geode deployments are
>going to be in the cloud, the servers are/going to be configured identically.
>Which means that resource management is a cluster level setting as opposed to
>a server level setting.
I think we need to remove the eviction and critical heap settings from the
{start server} command and instead create a new gfsh command
{configure resource-manager --eviction-heap-percentage=80
--critical-heap-percentage=95 }
> exporting cluster-config does not export resource-manager settings
> ------------------------------------------------------------------
>
> Key: GEODE-66
> URL: https://issues.apache.org/jira/browse/GEODE-66
> Project: Geode
> Issue Type: Bug
> Components: management
> Reporter: Swapnil Bawaskar
> Assignee: Jinmei Liao
> Priority: Minor
>
> When a server is started using gfsh with eviction and/or critical heap
> percentages specified, and we later do a export cluster-configuration, the
> eviction threshold is not reflected in the cluster config zip archive.
> To reproduce:
> 1. Start a locator and a server and create some regions:
> {code}
> gfsh>start locator --name=loc1
> gfsh>start server --name=server1 --eviction-heap-percentage=90
> gfsh>create region --name=usertable --type=PARTITION_REDUNDANT_HEAP_LRU
> {code}
> 2. export the cluster configuration:
> {code}
> gfsh>export cluster-configuration --zip-file-name=cluster-config.zip
> {code}
> 3. unzip cluster-config.zip to notice that cluster/cluster.xml does not a
> resource-manager element.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
