PAVEL DERENDYAEV created MESOS-8902:
---------------------------------------
Summary: Mesos masters stop to respond after zookeeper cluster
recover
Key: MESOS-8902
URL: https://issues.apache.org/jira/browse/MESOS-8902
Project: Mesos
Issue Type: Bug
Components: master
Affects Versions: 1.5.0, 1.4.1, 1.4.0
Reporter: PAVEL DERENDYAEV
Fix For: 1.3.0
Hi.
My setup consists of 3 zookeeper hosts and 3 mesos master hosts.
It's all fine being started. But if:
# Stop 2 zk nodes.
# Wait till all masters stop to respond to HTTP requests.
# Start 2 zk nodes.
# Then only leader respond to HTTP requests. The other two do not respond at
all and logged:
```
2018-05-10 14:41:16,274:1(0x7fb4989a7700):ZOO_WARN@zookeeper_interest@1597:
Exceeded deadline by 11ms
2018-05-10 14:41:16,274:1(0x7fb4737fe700):ZOO_WARN@zookeeper_interest@1597:
Exceeded deadline by 11ms
2018-05-10 14:42:56,377:1(0x7fb4737fe700):ZOO_WARN@zookeeper_interest@1597:
Exceeded deadline by 13ms
2018-05-10 14:42:56,377:1(0x7fb4989a7700):ZOO_WARN@zookeeper_interest@1597:
Exceeded deadline by 13ms
2018-05-10 14:43:19,783:1(0x7fb470ff9700):ZOO_WARN@zookeeper_interest@1597:
Exceeded deadline by 12ms
```
Being restarted these 2 master nodes starts to respond fine.
Reproduced on 1.4.0, 1.4.1, 1.5.0, do not reproduced on 1.3.0.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
