Details
-
Bug
-
Status: Closed
-
Critical
-
Resolution: Fixed
-
None
-
None
Description
Once Ignite persistence is enabled, the cluster has to be activated manually before any operation can be processed by it. That's needed to ensure that all the nodes that have data on disk have joined the cluster after a restart and that the topology is considered to be stable.
Some details can be found in this discussion:
- Usage section of this doc [1]. Look for a paragraph that starts with _"Next, when Ignite sees that the persistence is enabled it moves the cluster from active to inactive state making sure that applications cannot​ modify the data until allowed."_
- http://apache-ignite-developers.2346864.n4.nabble.com/Control-sh-script-and-cluster-activation-td20821.html
Document the manual cluster activation on a dedicated page [2] at least considering the following:
- Overview. Why it's required? Move the paragraph from [1] and expand if required.
- Show an exception that happens if you try to execute a SQL or key-value query or run a computation against an activated cluster.
- Explain how to activate the cluster with:
- From command line with control.sh script.
- From web console.
- From source code with Ignite.activate() method.
- Add a callout that the automatic cluster activation is to be added in the future. You may refer to this discussion [3] and a wiki page [4].
[1] https://apacheignite.readme.io/v2.2/docs/ignite-persistence-23#section-usage
[2] https://apacheignite.readme.io/v2.2/docs/cluster-activation
[3] http://apache-ignite-developers.2346864.n4.nabble.com/Cluster-auto-activation-design-proposal-td20295.html
[4] https://cwiki.apache.org/confluence/display/IGNITE/Automatic+activation+design+-+draft