[
https://issues.apache.org/jira/browse/GEODE-3107?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Fred Krone updated GEODE-3107:
------------------------------
Description:
Currently a large snapshot puts a lot of pressure on the locator which needs to
hold all keys and they do not get destroyed.
Snapshot goes through the JMX manager but snapshots can also be parallelized
and in theory this will probably improve performance.
It would be useful to benchmark the current gfsh export time on 1 TB region and
compare it with performance improvements from this ticket.
was:
Currently a large snapshot puts a lot of pressure on the locator which needs to
hold all keys.
Snapshot goes through the JMX manager but snapshots can also be parallelized
and in theory this will probably improve performance.
It would be useful to benchmark the current gfsh export time on 1 TB region and
compare it with performance improvements from this ticket.
> Improve gfsh export time when doing a snapshot
> ----------------------------------------------
>
> Key: GEODE-3107
> URL: https://issues.apache.org/jira/browse/GEODE-3107
> Project: Geode
> Issue Type: Bug
> Components: regions
> Affects Versions: 1.3.0
> Reporter: Fred Krone
>
> Currently a large snapshot puts a lot of pressure on the locator which needs
> to hold all keys and they do not get destroyed.
> Snapshot goes through the JMX manager but snapshots can also be parallelized
> and in theory this will probably improve performance.
> It would be useful to benchmark the current gfsh export time on 1 TB region
> and compare it with performance improvements from this ticket.
--
This message was sent by Atlassian JIRA
(v6.4.14#64029)
