You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@ignite.apache.org by "Ivan Daschinskiy (Jira)" <ji...@apache.org> on 2020/09/29 08:23:00 UTC
[jira] [Updated] (IGNITE-13495)
ZookeeperDiscoverySpiMBeanImpl#getCoordinator can return invalid node as
coordinator
[ https://issues.apache.org/jira/browse/IGNITE-13495?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Ivan Daschinskiy updated IGNITE-13495:
--------------------------------------
Description:
Due to invalid algorithm in {{org.apache.ignite.spi.discovery.zk.internal.ZookeeperDiscoveryImpl#getCoordinator}}
sometimes invalid coordinator could be return
Consider scenarion:
1. Start server #1
2. Start client
3. Start server #2
4. Stop server #1
After this, {{ZookeeperDiscoverySpiMBeanImpl#getCoordinator}} returns as coordinator a client, because it is the oldest node in topology.
We should fix {{org.apache.ignite.spi.discovery.zk.internal.ZookeeperDiscoveryImpl#getCoordinator}} to return *oldest server*, not any node.
was:
Due to invalid algorithm in {{org.apache.ignite.spi.discovery.zk.internal.ZookeeperDiscoveryImpl#getCoordinator}}
sometimes invalid coordinator could be return
Consider scenarion:
1. Start server #1
2. Start client
3. Start server #2
4. Stop server #1
After this, {{ZookeeperDiscoverySpiMBeanImpl#getCoordinator}} returns as coordinator a client, because it is the oldest node in topology.
We should rewrite {{org.apache.ignite.spi.discovery.zk.internal.ZookeeperDiscoveryImpl#getCoordinator}} to return *oldest server*, not any node.
> ZookeeperDiscoverySpiMBeanImpl#getCoordinator can return invalid node as coordinator
> ------------------------------------------------------------------------------------
>
> Key: IGNITE-13495
> URL: https://issues.apache.org/jira/browse/IGNITE-13495
> Project: Ignite
> Issue Type: Bug
> Affects Versions: 2.9
> Reporter: Ivan Daschinskiy
> Priority: Major
> Labels: newbie, zookeeper
> Fix For: 2.10
>
>
> Due to invalid algorithm in {{org.apache.ignite.spi.discovery.zk.internal.ZookeeperDiscoveryImpl#getCoordinator}}
> sometimes invalid coordinator could be return
> Consider scenarion:
> 1. Start server #1
> 2. Start client
> 3. Start server #2
> 4. Stop server #1
> After this, {{ZookeeperDiscoverySpiMBeanImpl#getCoordinator}} returns as coordinator a client, because it is the oldest node in topology.
> We should fix {{org.apache.ignite.spi.discovery.zk.internal.ZookeeperDiscoveryImpl#getCoordinator}} to return *oldest server*, not any node.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)