[
https://issues.apache.org/jira/browse/GEODE-5053?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16436040#comment-16436040
]
Barbara Pruijn commented on GEODE-5053:
---------------------------------------
[~pchandra] Here is the flow: if you start a cluster with brand new locators
and servers (with no configuration on them), you can import the cluster
configuration when the full cluster is up and running (locators and servers are
running). If the locators already have configuration and/or the servers have
configuration, the import command will fail.
Once you have imported the configuration on the running cluster, then you can
import the data. If the data that you import does not match the configuration,
this command will fail. (e.g. load data in a region that is not defined in the
configuration).
In the above scenario, you can import cluster config on running servers. But if
there is any configuration on any server or locator, you cannot import the
configuration.
https://issues.apache.org/jira/browse/GEODE-4893 fixed a bug related to this
and should be released in geode 1.6.0.
> Import and export should be supported with full cluster running
> ---------------------------------------------------------------
>
> Key: GEODE-5053
> URL: https://issues.apache.org/jira/browse/GEODE-5053
> Project: Geode
> Issue Type: Improvement
> Components: gfsh, management
> Affects Versions: 1.4.0, 1.5.0
> Reporter: Pulkit Chandra
> Priority: Major
>
> We want to provide support to export /import config and data. We dont have
> the ability to just run locators and not servers as we use BOSH to manage
> Geode, which is required for `import` today
> **Why** : As a operator if data contained in Geode is important, I would like
> to ensure I can back it up. Use cases range from ensuring your data doesnt
> get lost to migrating your data to another cluster.
> **How**: `Import` and `export` command should be the way to go
> **Limitations**: Cant have locators running and servers not running. Full
> cluster would be running before import is called. We can do rolling restart,
> if that is needed.
>
> its related to the issue https://issues.apache.org/jira/browse/GEODE-4893
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
