You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@ignite.apache.org by "Alexander Lapin (Jira)" <ji...@apache.org> on 2020/07/09 11:57:00 UTC
[jira] [Created] (IGNITE-13236) Rebalance doesn't happen after
restart if node failed at the moment of rebalance
Alexander Lapin created IGNITE-13236:
----------------------------------------
Summary: Rebalance doesn't happen after restart if node failed at the moment of rebalance
Key: IGNITE-13236
URL: https://issues.apache.org/jira/browse/IGNITE-13236
Project: Ignite
Issue Type: Bug
Environment: !Screen Shot 2019-08-05 at 18.20.46.png!
Reporter: Alexander Lapin
Attachments: Screen Shot 2019-08-05 at 18.20.46.png, image-2020-07-09-14-55-51-574.png
Steps to reproduce:
1. Start 2 nodes in baseline
2. Stream a lot of data.
3. Stop the second node, clean the persistence and start it again(with the same consistent id
4. Kill the first node at the moment of rebalancing (make sure that not all data was rebalanced)
5. Partition loss will be detected.
6. Return first node to the cluster.
7. Result:
!image-2020-07-09-14-55-51-574.png!
Data is not rebalanced, All data(primary and backup) is stored on the first node only, while the second node has the same small subset of the data. It's reproducible with any Partition loss policy. Resetting lost partitions, activation/deactivation doesn't help.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)