You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@zookeeper.apache.org by "Baohua (Jira)" <ji...@apache.org> on 2020/02/04 18:31:00 UTC
[jira] [Updated] (ZOOKEEPER-3717) Zookeeper won't work with swarm
service VIP when enabling healthcheck
[ https://issues.apache.org/jira/browse/ZOOKEEPER-3717?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Baohua updated ZOOKEEPER-3717:
------------------------------
Description:
*Issue*:
When use swarm service to create a zookeeper cluster with service VIP and enable healthcheck in the zookeeper docker image (even directly return 0), the client cannot connect to the server (status is connecting, and then retry), but the server status will report everything is OK.
*Root cause*:
When swarm service is started, the service VIP does not work until the healthcheck is passed. At that time, seems the client service is in an abnormal status that even after healthcheck is passed, and the zookeeper cluster is normal, no client can connect to the zookeeper.
*Solution (potential)*:
After the cluster status becomes OK, need to reset the client service to allow connection.
was:
Issue:
When use swarm service to create a zookeeper cluster with service VIP and enable healthcheck in the zookeeper docker image (even directly return 0), the client cannot connect to the server (status is connecting, and then retry), but the server status will report everything is OK.
Root cause:
When swarm service is started, the service VIP does not work until the healthcheck is passed. At that time, seems the client service is in an abnormal status that even after healthcheck is passed, and the zookeeper cluster is normal, no client can connect to the zookeeper.
Solution (potential):
After the cluster status becomes OK, need to reset the client service to allow connection.
> Zookeeper won't work with swarm service VIP when enabling healthcheck
> ---------------------------------------------------------------------
>
> Key: ZOOKEEPER-3717
> URL: https://issues.apache.org/jira/browse/ZOOKEEPER-3717
> Project: ZooKeeper
> Issue Type: Bug
> Affects Versions: 3.5.1, 3.5.2, 3.5.3, 3.5.4, 3.5.5, 3.4.14, 3.5.6
> Reporter: Baohua
> Priority: Major
>
> *Issue*:
> When use swarm service to create a zookeeper cluster with service VIP and enable healthcheck in the zookeeper docker image (even directly return 0), the client cannot connect to the server (status is connecting, and then retry), but the server status will report everything is OK.
> *Root cause*:
> When swarm service is started, the service VIP does not work until the healthcheck is passed. At that time, seems the client service is in an abnormal status that even after healthcheck is passed, and the zookeeper cluster is normal, no client can connect to the zookeeper.
> *Solution (potential)*:
> After the cluster status becomes OK, need to reset the client service to allow connection.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)