[jira] [Commented] (ZOOKEEPER-3756) Members failing to rejoin quorum
[
https://issues.apache.org/jira/browse/ZOOKEEPER-3756?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17111842#comment-17111842
]
Mate Szalay-Beko commented on ZOOKEEPER-3756:
-
Thanks [~dshi] for the tests and the feedback, it is great that the fix solved
your problem (and not an other problem :p )!
> Members failing to rejoin quorum
>
>
> Key: ZOOKEEPER-3756
> URL: https://issues.apache.org/jira/browse/ZOOKEEPER-3756
> Project: ZooKeeper
> Issue Type: Bug
> Components: leaderElection
>Affects Versions: 3.5.6, 3.5.7
>Reporter: Dai Shi
>Assignee: Mate Szalay-Beko
>Priority: Major
> Labels: pull-request-available
> Fix For: 3.6.1, 3.5.8
>
> Attachments: Dockerfile, configmap.yaml, docker-entrypoint.sh,
> jmx.yaml, zoo-0.log, zoo-1.log, zoo-2.log, zoo-service.yaml, zookeeper.yaml
>
> Time Spent: 3.5h
> Remaining Estimate: 0h
>
> Not sure if this is the place to ask, please close if it's not.
> I am seeing some behavior that I can't explain since upgrading to 3.5:
> In a 5 member quorum, when server 3 is the leader and each server has this in
> their configuration:
> {code:java}
> server.1=100.71.255.254:2888:3888:participant;2181
> server.2=100.71.255.253:2888:3888:participant;2181
> server.3=100.71.255.252:2888:3888:participant;2181
> server.4=100.71.255.251:2888:3888:participant;2181
> server.5=100.71.255.250:2888:3888:participant;2181{code}
> If servers 1 or 2 are restarted, they fail to rejoin the quorum with this in
> the logs:
> {code:java}
> 2020-03-11 20:23:35,720 [myid:2] - INFO
> [QuorumPeer[myid=2](plain=0.0.0.0:2181)(secure=disabled):QuorumPeer@1175] -
> LOOKING
> 2020-03-11 20:23:35,721 [myid:2] - INFO
> [QuorumPeer[myid=2](plain=0.0.0.0:2181)(secure=disabled):FastLeaderElection@885]
> - New election. My id = 2, proposed zxid=0x1b8005f4bba
> 2020-03-11 20:23:35,733 [myid:2] - INFO
> [WorkerSender[myid=2]:QuorumCnxManager@438] - Have smaller server identifier,
> so dropping the connection: (3, 2)
> 2020-03-11 20:23:35,734 [myid:2] - INFO
> [0.0.0.0/0.0.0.0:3888:QuorumCnxManager$Listener@924] - Received connection
> request 100.126.116.201:36140
> 2020-03-11 20:23:35,735 [myid:2] - INFO
> [WorkerSender[myid=2]:QuorumCnxManager@438] - Have smaller server identifier,
> so dropping the connection: (4, 2)
> 2020-03-11 20:23:35,740 [myid:2] - INFO
> [WorkerSender[myid=2]:QuorumCnxManager@438] - Have smaller server identifier,
> so dropping the connection: (5, 2)
> 2020-03-11 20:23:35,740 [myid:2] - INFO
> [0.0.0.0/0.0.0.0:3888:QuorumCnxManager$Listener@924] - Received connection
> request 100.126.116.201:36142
> 2020-03-11 20:23:35,740 [myid:2] - INFO
> [WorkerReceiver[myid=2]:FastLeaderElection@679] - Notification: 2 (message
> format version), 2 (n.leader), 0x1b8005f4bba (n.zxid), 0x1 (n.round), LOOKING
> (n.state), 2 (n.sid), 0x1b8 (n.peerEPoch), LOOKING (my state)0 (n.config
> version)
> 2020-03-11 20:23:35,742 [myid:2] - WARN
> [SendWorker:3:QuorumCnxManager$SendWorker@1143] - Interrupted while waiting
> for message on queue
> java.lang.InterruptedException
> at
> java.util.concurrent.locks.AbstractQueuedSynchronizer$ConditionObject.reportInterruptAfterWait(AbstractQueuedSynchronizer.java:2014)
> at
> java.util.concurrent.locks.AbstractQueuedSynchronizer$ConditionObject.awaitNanos(AbstractQueuedSynchronizer.java:2088)
> at
> java.util.concurrent.ArrayBlockingQueue.poll(ArrayBlockingQueue.java:418)
> at
> org.apache.zookeeper.server.quorum.QuorumCnxManager.pollSendQueue(QuorumCnxManager.java:1294)
> at
> org.apache.zookeeper.server.quorum.QuorumCnxManager.access$700(QuorumCnxManager.java:82)
> at
> org.apache.zookeeper.server.quorum.QuorumCnxManager$SendWorker.run(QuorumCnxManager.java:1131)
> 2020-03-11 20:23:35,744 [myid:2] - WARN
> [SendWorker:3:QuorumCnxManager$SendWorker@1153] - Send worker leaving thread
> id 3 my id = 2
> 2020-03-11 20:23:35,745 [myid:2] - WARN
> [RecvWorker:3:QuorumCnxManager$RecvWorker@1230] - Interrupting
> SendWorker{code}
> The only way I can seem to get them to rejoin the quorum is to restart the
> leader.
> However, if I remove server 4 and 5 from the configuration of server 1 or 2
> (so only servers 1, 2, and 3 remain in the configuration file), then they can
> rejoin the quorum fine. Is this expected and am I doing something wrong? Any
> help or explanation would be greatly appreciated. Thank you.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
[jira] [Commented] (ZOOKEEPER-3756) Members failing to rejoin quorum
[
https://issues.apache.org/jira/browse/ZOOKEEPER-3756?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17111589#comment-17111589
]
Dai Shi commented on ZOOKEEPER-3756:
Hi Mate,
I just wanted to report back after testing 3.5.8. I am happy to say that it
seems to work well after a brief bit of testing. I am no longer setting
`-Dzookeeper.cnxTimeout=500`, and now when I roll the leader the cluster
downtime is only 2-3 seconds instead of 30+ seconds.
Thanks again for helping me debug and creating this fix!
> Members failing to rejoin quorum
>
>
> Key: ZOOKEEPER-3756
> URL: https://issues.apache.org/jira/browse/ZOOKEEPER-3756
> Project: ZooKeeper
> Issue Type: Bug
> Components: leaderElection
>Affects Versions: 3.5.6, 3.5.7
>Reporter: Dai Shi
>Assignee: Mate Szalay-Beko
>Priority: Major
> Labels: pull-request-available
> Fix For: 3.6.1, 3.5.8
>
> Attachments: Dockerfile, configmap.yaml, docker-entrypoint.sh,
> jmx.yaml, zoo-0.log, zoo-1.log, zoo-2.log, zoo-service.yaml, zookeeper.yaml
>
> Time Spent: 3.5h
> Remaining Estimate: 0h
>
> Not sure if this is the place to ask, please close if it's not.
> I am seeing some behavior that I can't explain since upgrading to 3.5:
> In a 5 member quorum, when server 3 is the leader and each server has this in
> their configuration:
> {code:java}
> server.1=100.71.255.254:2888:3888:participant;2181
> server.2=100.71.255.253:2888:3888:participant;2181
> server.3=100.71.255.252:2888:3888:participant;2181
> server.4=100.71.255.251:2888:3888:participant;2181
> server.5=100.71.255.250:2888:3888:participant;2181{code}
> If servers 1 or 2 are restarted, they fail to rejoin the quorum with this in
> the logs:
> {code:java}
> 2020-03-11 20:23:35,720 [myid:2] - INFO
> [QuorumPeer[myid=2](plain=0.0.0.0:2181)(secure=disabled):QuorumPeer@1175] -
> LOOKING
> 2020-03-11 20:23:35,721 [myid:2] - INFO
> [QuorumPeer[myid=2](plain=0.0.0.0:2181)(secure=disabled):FastLeaderElection@885]
> - New election. My id = 2, proposed zxid=0x1b8005f4bba
> 2020-03-11 20:23:35,733 [myid:2] - INFO
> [WorkerSender[myid=2]:QuorumCnxManager@438] - Have smaller server identifier,
> so dropping the connection: (3, 2)
> 2020-03-11 20:23:35,734 [myid:2] - INFO
> [0.0.0.0/0.0.0.0:3888:QuorumCnxManager$Listener@924] - Received connection
> request 100.126.116.201:36140
> 2020-03-11 20:23:35,735 [myid:2] - INFO
> [WorkerSender[myid=2]:QuorumCnxManager@438] - Have smaller server identifier,
> so dropping the connection: (4, 2)
> 2020-03-11 20:23:35,740 [myid:2] - INFO
> [WorkerSender[myid=2]:QuorumCnxManager@438] - Have smaller server identifier,
> so dropping the connection: (5, 2)
> 2020-03-11 20:23:35,740 [myid:2] - INFO
> [0.0.0.0/0.0.0.0:3888:QuorumCnxManager$Listener@924] - Received connection
> request 100.126.116.201:36142
> 2020-03-11 20:23:35,740 [myid:2] - INFO
> [WorkerReceiver[myid=2]:FastLeaderElection@679] - Notification: 2 (message
> format version), 2 (n.leader), 0x1b8005f4bba (n.zxid), 0x1 (n.round), LOOKING
> (n.state), 2 (n.sid), 0x1b8 (n.peerEPoch), LOOKING (my state)0 (n.config
> version)
> 2020-03-11 20:23:35,742 [myid:2] - WARN
> [SendWorker:3:QuorumCnxManager$SendWorker@1143] - Interrupted while waiting
> for message on queue
> java.lang.InterruptedException
> at
> java.util.concurrent.locks.AbstractQueuedSynchronizer$ConditionObject.reportInterruptAfterWait(AbstractQueuedSynchronizer.java:2014)
> at
> java.util.concurrent.locks.AbstractQueuedSynchronizer$ConditionObject.awaitNanos(AbstractQueuedSynchronizer.java:2088)
> at
> java.util.concurrent.ArrayBlockingQueue.poll(ArrayBlockingQueue.java:418)
> at
> org.apache.zookeeper.server.quorum.QuorumCnxManager.pollSendQueue(QuorumCnxManager.java:1294)
> at
> org.apache.zookeeper.server.quorum.QuorumCnxManager.access$700(QuorumCnxManager.java:82)
> at
> org.apache.zookeeper.server.quorum.QuorumCnxManager$SendWorker.run(QuorumCnxManager.java:1131)
> 2020-03-11 20:23:35,744 [myid:2] - WARN
> [SendWorker:3:QuorumCnxManager$SendWorker@1153] - Send worker leaving thread
> id 3 my id = 2
> 2020-03-11 20:23:35,745 [myid:2] - WARN
> [RecvWorker:3:QuorumCnxManager$RecvWorker@1230] - Interrupting
> SendWorker{code}
> The only way I can seem to get them to rejoin the quorum is to restart the
> leader.
> However, if I remove server 4 and 5 from the configuration of server 1 or 2
> (so only servers 1, 2, and 3 remain in the configuration file), then they can
> rejoin the quorum fine. Is this expected and am I doing something wrong? Any
> help or explanation would be greatly appreciated. Thank you.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
[jira] [Commented] (ZOOKEEPER-3756) Members failing to rejoin quorum
[
https://issues.apache.org/jira/browse/ZOOKEEPER-3756?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17093880#comment-17093880
]
Dai Shi commented on ZOOKEEPER-3756:
I will definitely test it once it's available. I will join the dev zookeeper
list to keep an eye on it.
> Members failing to rejoin quorum
>
>
> Key: ZOOKEEPER-3756
> URL: https://issues.apache.org/jira/browse/ZOOKEEPER-3756
> Project: ZooKeeper
> Issue Type: Improvement
> Components: leaderElection
>Affects Versions: 3.5.6, 3.5.7
>Reporter: Dai Shi
>Assignee: Mate Szalay-Beko
>Priority: Major
> Labels: pull-request-available
> Fix For: 3.6.1, 3.5.8
>
> Attachments: Dockerfile, configmap.yaml, docker-entrypoint.sh,
> jmx.yaml, zoo-0.log, zoo-1.log, zoo-2.log, zoo-service.yaml, zookeeper.yaml
>
> Time Spent: 3.5h
> Remaining Estimate: 0h
>
> Not sure if this is the place to ask, please close if it's not.
> I am seeing some behavior that I can't explain since upgrading to 3.5:
> In a 5 member quorum, when server 3 is the leader and each server has this in
> their configuration:
> {code:java}
> server.1=100.71.255.254:2888:3888:participant;2181
> server.2=100.71.255.253:2888:3888:participant;2181
> server.3=100.71.255.252:2888:3888:participant;2181
> server.4=100.71.255.251:2888:3888:participant;2181
> server.5=100.71.255.250:2888:3888:participant;2181{code}
> If servers 1 or 2 are restarted, they fail to rejoin the quorum with this in
> the logs:
> {code:java}
> 2020-03-11 20:23:35,720 [myid:2] - INFO
> [QuorumPeer[myid=2](plain=0.0.0.0:2181)(secure=disabled):QuorumPeer@1175] -
> LOOKING
> 2020-03-11 20:23:35,721 [myid:2] - INFO
> [QuorumPeer[myid=2](plain=0.0.0.0:2181)(secure=disabled):FastLeaderElection@885]
> - New election. My id = 2, proposed zxid=0x1b8005f4bba
> 2020-03-11 20:23:35,733 [myid:2] - INFO
> [WorkerSender[myid=2]:QuorumCnxManager@438] - Have smaller server identifier,
> so dropping the connection: (3, 2)
> 2020-03-11 20:23:35,734 [myid:2] - INFO
> [0.0.0.0/0.0.0.0:3888:QuorumCnxManager$Listener@924] - Received connection
> request 100.126.116.201:36140
> 2020-03-11 20:23:35,735 [myid:2] - INFO
> [WorkerSender[myid=2]:QuorumCnxManager@438] - Have smaller server identifier,
> so dropping the connection: (4, 2)
> 2020-03-11 20:23:35,740 [myid:2] - INFO
> [WorkerSender[myid=2]:QuorumCnxManager@438] - Have smaller server identifier,
> so dropping the connection: (5, 2)
> 2020-03-11 20:23:35,740 [myid:2] - INFO
> [0.0.0.0/0.0.0.0:3888:QuorumCnxManager$Listener@924] - Received connection
> request 100.126.116.201:36142
> 2020-03-11 20:23:35,740 [myid:2] - INFO
> [WorkerReceiver[myid=2]:FastLeaderElection@679] - Notification: 2 (message
> format version), 2 (n.leader), 0x1b8005f4bba (n.zxid), 0x1 (n.round), LOOKING
> (n.state), 2 (n.sid), 0x1b8 (n.peerEPoch), LOOKING (my state)0 (n.config
> version)
> 2020-03-11 20:23:35,742 [myid:2] - WARN
> [SendWorker:3:QuorumCnxManager$SendWorker@1143] - Interrupted while waiting
> for message on queue
> java.lang.InterruptedException
> at
> java.util.concurrent.locks.AbstractQueuedSynchronizer$ConditionObject.reportInterruptAfterWait(AbstractQueuedSynchronizer.java:2014)
> at
> java.util.concurrent.locks.AbstractQueuedSynchronizer$ConditionObject.awaitNanos(AbstractQueuedSynchronizer.java:2088)
> at
> java.util.concurrent.ArrayBlockingQueue.poll(ArrayBlockingQueue.java:418)
> at
> org.apache.zookeeper.server.quorum.QuorumCnxManager.pollSendQueue(QuorumCnxManager.java:1294)
> at
> org.apache.zookeeper.server.quorum.QuorumCnxManager.access$700(QuorumCnxManager.java:82)
> at
> org.apache.zookeeper.server.quorum.QuorumCnxManager$SendWorker.run(QuorumCnxManager.java:1131)
> 2020-03-11 20:23:35,744 [myid:2] - WARN
> [SendWorker:3:QuorumCnxManager$SendWorker@1153] - Send worker leaving thread
> id 3 my id = 2
> 2020-03-11 20:23:35,745 [myid:2] - WARN
> [RecvWorker:3:QuorumCnxManager$RecvWorker@1230] - Interrupting
> SendWorker{code}
> The only way I can seem to get them to rejoin the quorum is to restart the
> leader.
> However, if I remove server 4 and 5 from the configuration of server 1 or 2
> (so only servers 1, 2, and 3 remain in the configuration file), then they can
> rejoin the quorum fine. Is this expected and am I doing something wrong? Any
> help or explanation would be greatly appreciated. Thank you.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
[jira] [Commented] (ZOOKEEPER-3756) Members failing to rejoin quorum
[
https://issues.apache.org/jira/browse/ZOOKEEPER-3756?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17093144#comment-17093144
]
Mate Szalay-Beko commented on ZOOKEEPER-3756:
-
I think I will drive the 3.5.8 release. I hope I will have time to make the
first release candidate on this week, or early next week. I will send out the
official announcement on the dev mailing list. I would appreciate if you can
test it.
Are you on the dev zookeeper list? Or should I forward the RC to you?
https://zookeeper.apache.org/lists.html
> Members failing to rejoin quorum
>
>
> Key: ZOOKEEPER-3756
> URL: https://issues.apache.org/jira/browse/ZOOKEEPER-3756
> Project: ZooKeeper
> Issue Type: Improvement
> Components: leaderElection
>Affects Versions: 3.5.6, 3.5.7
>Reporter: Dai Shi
>Assignee: Mate Szalay-Beko
>Priority: Major
> Labels: pull-request-available
> Fix For: 3.6.1, 3.5.8
>
> Attachments: Dockerfile, configmap.yaml, docker-entrypoint.sh,
> jmx.yaml, zoo-0.log, zoo-1.log, zoo-2.log, zoo-service.yaml, zookeeper.yaml
>
> Time Spent: 3.5h
> Remaining Estimate: 0h
>
> Not sure if this is the place to ask, please close if it's not.
> I am seeing some behavior that I can't explain since upgrading to 3.5:
> In a 5 member quorum, when server 3 is the leader and each server has this in
> their configuration:
> {code:java}
> server.1=100.71.255.254:2888:3888:participant;2181
> server.2=100.71.255.253:2888:3888:participant;2181
> server.3=100.71.255.252:2888:3888:participant;2181
> server.4=100.71.255.251:2888:3888:participant;2181
> server.5=100.71.255.250:2888:3888:participant;2181{code}
> If servers 1 or 2 are restarted, they fail to rejoin the quorum with this in
> the logs:
> {code:java}
> 2020-03-11 20:23:35,720 [myid:2] - INFO
> [QuorumPeer[myid=2](plain=0.0.0.0:2181)(secure=disabled):QuorumPeer@1175] -
> LOOKING
> 2020-03-11 20:23:35,721 [myid:2] - INFO
> [QuorumPeer[myid=2](plain=0.0.0.0:2181)(secure=disabled):FastLeaderElection@885]
> - New election. My id = 2, proposed zxid=0x1b8005f4bba
> 2020-03-11 20:23:35,733 [myid:2] - INFO
> [WorkerSender[myid=2]:QuorumCnxManager@438] - Have smaller server identifier,
> so dropping the connection: (3, 2)
> 2020-03-11 20:23:35,734 [myid:2] - INFO
> [0.0.0.0/0.0.0.0:3888:QuorumCnxManager$Listener@924] - Received connection
> request 100.126.116.201:36140
> 2020-03-11 20:23:35,735 [myid:2] - INFO
> [WorkerSender[myid=2]:QuorumCnxManager@438] - Have smaller server identifier,
> so dropping the connection: (4, 2)
> 2020-03-11 20:23:35,740 [myid:2] - INFO
> [WorkerSender[myid=2]:QuorumCnxManager@438] - Have smaller server identifier,
> so dropping the connection: (5, 2)
> 2020-03-11 20:23:35,740 [myid:2] - INFO
> [0.0.0.0/0.0.0.0:3888:QuorumCnxManager$Listener@924] - Received connection
> request 100.126.116.201:36142
> 2020-03-11 20:23:35,740 [myid:2] - INFO
> [WorkerReceiver[myid=2]:FastLeaderElection@679] - Notification: 2 (message
> format version), 2 (n.leader), 0x1b8005f4bba (n.zxid), 0x1 (n.round), LOOKING
> (n.state), 2 (n.sid), 0x1b8 (n.peerEPoch), LOOKING (my state)0 (n.config
> version)
> 2020-03-11 20:23:35,742 [myid:2] - WARN
> [SendWorker:3:QuorumCnxManager$SendWorker@1143] - Interrupted while waiting
> for message on queue
> java.lang.InterruptedException
> at
> java.util.concurrent.locks.AbstractQueuedSynchronizer$ConditionObject.reportInterruptAfterWait(AbstractQueuedSynchronizer.java:2014)
> at
> java.util.concurrent.locks.AbstractQueuedSynchronizer$ConditionObject.awaitNanos(AbstractQueuedSynchronizer.java:2088)
> at
> java.util.concurrent.ArrayBlockingQueue.poll(ArrayBlockingQueue.java:418)
> at
> org.apache.zookeeper.server.quorum.QuorumCnxManager.pollSendQueue(QuorumCnxManager.java:1294)
> at
> org.apache.zookeeper.server.quorum.QuorumCnxManager.access$700(QuorumCnxManager.java:82)
> at
> org.apache.zookeeper.server.quorum.QuorumCnxManager$SendWorker.run(QuorumCnxManager.java:1131)
> 2020-03-11 20:23:35,744 [myid:2] - WARN
> [SendWorker:3:QuorumCnxManager$SendWorker@1153] - Send worker leaving thread
> id 3 my id = 2
> 2020-03-11 20:23:35,745 [myid:2] - WARN
> [RecvWorker:3:QuorumCnxManager$RecvWorker@1230] - Interrupting
> SendWorker{code}
> The only way I can seem to get them to rejoin the quorum is to restart the
> leader.
> However, if I remove server 4 and 5 from the configuration of server 1 or 2
> (so only servers 1, 2, and 3 remain in the configuration file), then they can
> rejoin the quorum fine. Is this expected and am I doing something wrong? Any
> help or explanation would be greatly appreciated. Thank you.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
[jira] [Commented] (ZOOKEEPER-3756) Members failing to rejoin quorum
[
https://issues.apache.org/jira/browse/ZOOKEEPER-3756?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17091945#comment-17091945
]
Dai Shi commented on ZOOKEEPER-3756:
Hi Mate,
Is there an estimated release time for 3.5.8? I'm hoping to try out the fix
soon, :)
> Members failing to rejoin quorum
>
>
> Key: ZOOKEEPER-3756
> URL: https://issues.apache.org/jira/browse/ZOOKEEPER-3756
> Project: ZooKeeper
> Issue Type: Improvement
> Components: leaderElection
>Affects Versions: 3.5.6, 3.5.7
>Reporter: Dai Shi
>Assignee: Mate Szalay-Beko
>Priority: Major
> Labels: pull-request-available
> Fix For: 3.6.1, 3.5.8
>
> Attachments: Dockerfile, configmap.yaml, docker-entrypoint.sh,
> jmx.yaml, zoo-0.log, zoo-1.log, zoo-2.log, zoo-service.yaml, zookeeper.yaml
>
> Time Spent: 3.5h
> Remaining Estimate: 0h
>
> Not sure if this is the place to ask, please close if it's not.
> I am seeing some behavior that I can't explain since upgrading to 3.5:
> In a 5 member quorum, when server 3 is the leader and each server has this in
> their configuration:
> {code:java}
> server.1=100.71.255.254:2888:3888:participant;2181
> server.2=100.71.255.253:2888:3888:participant;2181
> server.3=100.71.255.252:2888:3888:participant;2181
> server.4=100.71.255.251:2888:3888:participant;2181
> server.5=100.71.255.250:2888:3888:participant;2181{code}
> If servers 1 or 2 are restarted, they fail to rejoin the quorum with this in
> the logs:
> {code:java}
> 2020-03-11 20:23:35,720 [myid:2] - INFO
> [QuorumPeer[myid=2](plain=0.0.0.0:2181)(secure=disabled):QuorumPeer@1175] -
> LOOKING
> 2020-03-11 20:23:35,721 [myid:2] - INFO
> [QuorumPeer[myid=2](plain=0.0.0.0:2181)(secure=disabled):FastLeaderElection@885]
> - New election. My id = 2, proposed zxid=0x1b8005f4bba
> 2020-03-11 20:23:35,733 [myid:2] - INFO
> [WorkerSender[myid=2]:QuorumCnxManager@438] - Have smaller server identifier,
> so dropping the connection: (3, 2)
> 2020-03-11 20:23:35,734 [myid:2] - INFO
> [0.0.0.0/0.0.0.0:3888:QuorumCnxManager$Listener@924] - Received connection
> request 100.126.116.201:36140
> 2020-03-11 20:23:35,735 [myid:2] - INFO
> [WorkerSender[myid=2]:QuorumCnxManager@438] - Have smaller server identifier,
> so dropping the connection: (4, 2)
> 2020-03-11 20:23:35,740 [myid:2] - INFO
> [WorkerSender[myid=2]:QuorumCnxManager@438] - Have smaller server identifier,
> so dropping the connection: (5, 2)
> 2020-03-11 20:23:35,740 [myid:2] - INFO
> [0.0.0.0/0.0.0.0:3888:QuorumCnxManager$Listener@924] - Received connection
> request 100.126.116.201:36142
> 2020-03-11 20:23:35,740 [myid:2] - INFO
> [WorkerReceiver[myid=2]:FastLeaderElection@679] - Notification: 2 (message
> format version), 2 (n.leader), 0x1b8005f4bba (n.zxid), 0x1 (n.round), LOOKING
> (n.state), 2 (n.sid), 0x1b8 (n.peerEPoch), LOOKING (my state)0 (n.config
> version)
> 2020-03-11 20:23:35,742 [myid:2] - WARN
> [SendWorker:3:QuorumCnxManager$SendWorker@1143] - Interrupted while waiting
> for message on queue
> java.lang.InterruptedException
> at
> java.util.concurrent.locks.AbstractQueuedSynchronizer$ConditionObject.reportInterruptAfterWait(AbstractQueuedSynchronizer.java:2014)
> at
> java.util.concurrent.locks.AbstractQueuedSynchronizer$ConditionObject.awaitNanos(AbstractQueuedSynchronizer.java:2088)
> at
> java.util.concurrent.ArrayBlockingQueue.poll(ArrayBlockingQueue.java:418)
> at
> org.apache.zookeeper.server.quorum.QuorumCnxManager.pollSendQueue(QuorumCnxManager.java:1294)
> at
> org.apache.zookeeper.server.quorum.QuorumCnxManager.access$700(QuorumCnxManager.java:82)
> at
> org.apache.zookeeper.server.quorum.QuorumCnxManager$SendWorker.run(QuorumCnxManager.java:1131)
> 2020-03-11 20:23:35,744 [myid:2] - WARN
> [SendWorker:3:QuorumCnxManager$SendWorker@1153] - Send worker leaving thread
> id 3 my id = 2
> 2020-03-11 20:23:35,745 [myid:2] - WARN
> [RecvWorker:3:QuorumCnxManager$RecvWorker@1230] - Interrupting
> SendWorker{code}
> The only way I can seem to get them to rejoin the quorum is to restart the
> leader.
> However, if I remove server 4 and 5 from the configuration of server 1 or 2
> (so only servers 1, 2, and 3 remain in the configuration file), then they can
> rejoin the quorum fine. Is this expected and am I doing something wrong? Any
> help or explanation would be greatly appreciated. Thank you.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
[jira] [Commented] (ZOOKEEPER-3756) Members failing to rejoin quorum
[
https://issues.apache.org/jira/browse/ZOOKEEPER-3756?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17066127#comment-17066127
]
Lasaro Camargos commented on ZOOKEEPER-3756:
Thanks for the feedback. I've opened ZOOKEEPER-3769 with a slightly different
scenario but problematic in the same sense. To give the complete answer, I am
not using 0.0.0.0 addresses (not explicitly, at least) and not using containers.
[~symat], I appreciate your willingness to look into it. It's been troubling me
for some time.
> Members failing to rejoin quorum
>
>
> Key: ZOOKEEPER-3756
> URL: https://issues.apache.org/jira/browse/ZOOKEEPER-3756
> Project: ZooKeeper
> Issue Type: Improvement
> Components: leaderElection
>Affects Versions: 3.5.6, 3.5.7
>Reporter: Dai Shi
>Assignee: Mate Szalay-Beko
>Priority: Major
> Labels: pull-request-available
> Fix For: 3.6.1, 3.5.8
>
> Attachments: Dockerfile, configmap.yaml, docker-entrypoint.sh,
> jmx.yaml, zoo-0.log, zoo-1.log, zoo-2.log, zoo-service.yaml, zookeeper.yaml
>
> Time Spent: 3h
> Remaining Estimate: 0h
>
> Not sure if this is the place to ask, please close if it's not.
> I am seeing some behavior that I can't explain since upgrading to 3.5:
> In a 5 member quorum, when server 3 is the leader and each server has this in
> their configuration:
> {code:java}
> server.1=100.71.255.254:2888:3888:participant;2181
> server.2=100.71.255.253:2888:3888:participant;2181
> server.3=100.71.255.252:2888:3888:participant;2181
> server.4=100.71.255.251:2888:3888:participant;2181
> server.5=100.71.255.250:2888:3888:participant;2181{code}
> If servers 1 or 2 are restarted, they fail to rejoin the quorum with this in
> the logs:
> {code:java}
> 2020-03-11 20:23:35,720 [myid:2] - INFO
> [QuorumPeer[myid=2](plain=0.0.0.0:2181)(secure=disabled):QuorumPeer@1175] -
> LOOKING
> 2020-03-11 20:23:35,721 [myid:2] - INFO
> [QuorumPeer[myid=2](plain=0.0.0.0:2181)(secure=disabled):FastLeaderElection@885]
> - New election. My id = 2, proposed zxid=0x1b8005f4bba
> 2020-03-11 20:23:35,733 [myid:2] - INFO
> [WorkerSender[myid=2]:QuorumCnxManager@438] - Have smaller server identifier,
> so dropping the connection: (3, 2)
> 2020-03-11 20:23:35,734 [myid:2] - INFO
> [0.0.0.0/0.0.0.0:3888:QuorumCnxManager$Listener@924] - Received connection
> request 100.126.116.201:36140
> 2020-03-11 20:23:35,735 [myid:2] - INFO
> [WorkerSender[myid=2]:QuorumCnxManager@438] - Have smaller server identifier,
> so dropping the connection: (4, 2)
> 2020-03-11 20:23:35,740 [myid:2] - INFO
> [WorkerSender[myid=2]:QuorumCnxManager@438] - Have smaller server identifier,
> so dropping the connection: (5, 2)
> 2020-03-11 20:23:35,740 [myid:2] - INFO
> [0.0.0.0/0.0.0.0:3888:QuorumCnxManager$Listener@924] - Received connection
> request 100.126.116.201:36142
> 2020-03-11 20:23:35,740 [myid:2] - INFO
> [WorkerReceiver[myid=2]:FastLeaderElection@679] - Notification: 2 (message
> format version), 2 (n.leader), 0x1b8005f4bba (n.zxid), 0x1 (n.round), LOOKING
> (n.state), 2 (n.sid), 0x1b8 (n.peerEPoch), LOOKING (my state)0 (n.config
> version)
> 2020-03-11 20:23:35,742 [myid:2] - WARN
> [SendWorker:3:QuorumCnxManager$SendWorker@1143] - Interrupted while waiting
> for message on queue
> java.lang.InterruptedException
> at
> java.util.concurrent.locks.AbstractQueuedSynchronizer$ConditionObject.reportInterruptAfterWait(AbstractQueuedSynchronizer.java:2014)
> at
> java.util.concurrent.locks.AbstractQueuedSynchronizer$ConditionObject.awaitNanos(AbstractQueuedSynchronizer.java:2088)
> at
> java.util.concurrent.ArrayBlockingQueue.poll(ArrayBlockingQueue.java:418)
> at
> org.apache.zookeeper.server.quorum.QuorumCnxManager.pollSendQueue(QuorumCnxManager.java:1294)
> at
> org.apache.zookeeper.server.quorum.QuorumCnxManager.access$700(QuorumCnxManager.java:82)
> at
> org.apache.zookeeper.server.quorum.QuorumCnxManager$SendWorker.run(QuorumCnxManager.java:1131)
> 2020-03-11 20:23:35,744 [myid:2] - WARN
> [SendWorker:3:QuorumCnxManager$SendWorker@1153] - Send worker leaving thread
> id 3 my id = 2
> 2020-03-11 20:23:35,745 [myid:2] - WARN
> [RecvWorker:3:QuorumCnxManager$RecvWorker@1230] - Interrupting
> SendWorker{code}
> The only way I can seem to get them to rejoin the quorum is to restart the
> leader.
> However, if I remove server 4 and 5 from the configuration of server 1 or 2
> (so only servers 1, 2, and 3 remain in the configuration file), then they can
> rejoin the quorum fine. Is this expected and am I doing something wrong? Any
> help or explanation would be greatly appreciated. Thank you.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
[jira] [Commented] (ZOOKEEPER-3756) Members failing to rejoin quorum
[
https://issues.apache.org/jira/browse/ZOOKEEPER-3756?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17065412#comment-17065412
]
Mate Szalay-Beko commented on ZOOKEEPER-3756:
-
[~lasaro] this is a strange issue indeed. It seems to be a bit different from
this particular Jira. You can try the {{-Dzookeeper.cnxTimeout=500}} workaround
(it helped in the case we found with this Jira).
If it doesn't help, then I recommend to open an other Jira which I will be
happy to assign to myself. Full logs and config files would help a lot for
sure. Also the ZooKeeper and the Java version is important. And we had several
docker / kubernetes specific issues lately, so it is also important to know if
you are facing with this problem in Docker / Kubernetes.
This thread can be also interesting for you, especially if you use 0.0.0.0 in
your configs:
https://mail-archives.apache.org/mod_mbox/zookeeper-user/202003.mbox/%3CCAAMoRKLDVeL0jfEJndyP3pnVVxztthqZ35d_UM%2Bhurx7%3DqO_PQ%40mail.gmail.com%3E
> Members failing to rejoin quorum
>
>
> Key: ZOOKEEPER-3756
> URL: https://issues.apache.org/jira/browse/ZOOKEEPER-3756
> Project: ZooKeeper
> Issue Type: Improvement
> Components: leaderElection
>Affects Versions: 3.5.6, 3.5.7
>Reporter: Dai Shi
>Assignee: Mate Szalay-Beko
>Priority: Major
> Labels: pull-request-available
> Fix For: 3.6.1, 3.5.8
>
> Attachments: Dockerfile, configmap.yaml, docker-entrypoint.sh,
> jmx.yaml, zoo-0.log, zoo-1.log, zoo-2.log, zoo-service.yaml, zookeeper.yaml
>
> Time Spent: 3h
> Remaining Estimate: 0h
>
> Not sure if this is the place to ask, please close if it's not.
> I am seeing some behavior that I can't explain since upgrading to 3.5:
> In a 5 member quorum, when server 3 is the leader and each server has this in
> their configuration:
> {code:java}
> server.1=100.71.255.254:2888:3888:participant;2181
> server.2=100.71.255.253:2888:3888:participant;2181
> server.3=100.71.255.252:2888:3888:participant;2181
> server.4=100.71.255.251:2888:3888:participant;2181
> server.5=100.71.255.250:2888:3888:participant;2181{code}
> If servers 1 or 2 are restarted, they fail to rejoin the quorum with this in
> the logs:
> {code:java}
> 2020-03-11 20:23:35,720 [myid:2] - INFO
> [QuorumPeer[myid=2](plain=0.0.0.0:2181)(secure=disabled):QuorumPeer@1175] -
> LOOKING
> 2020-03-11 20:23:35,721 [myid:2] - INFO
> [QuorumPeer[myid=2](plain=0.0.0.0:2181)(secure=disabled):FastLeaderElection@885]
> - New election. My id = 2, proposed zxid=0x1b8005f4bba
> 2020-03-11 20:23:35,733 [myid:2] - INFO
> [WorkerSender[myid=2]:QuorumCnxManager@438] - Have smaller server identifier,
> so dropping the connection: (3, 2)
> 2020-03-11 20:23:35,734 [myid:2] - INFO
> [0.0.0.0/0.0.0.0:3888:QuorumCnxManager$Listener@924] - Received connection
> request 100.126.116.201:36140
> 2020-03-11 20:23:35,735 [myid:2] - INFO
> [WorkerSender[myid=2]:QuorumCnxManager@438] - Have smaller server identifier,
> so dropping the connection: (4, 2)
> 2020-03-11 20:23:35,740 [myid:2] - INFO
> [WorkerSender[myid=2]:QuorumCnxManager@438] - Have smaller server identifier,
> so dropping the connection: (5, 2)
> 2020-03-11 20:23:35,740 [myid:2] - INFO
> [0.0.0.0/0.0.0.0:3888:QuorumCnxManager$Listener@924] - Received connection
> request 100.126.116.201:36142
> 2020-03-11 20:23:35,740 [myid:2] - INFO
> [WorkerReceiver[myid=2]:FastLeaderElection@679] - Notification: 2 (message
> format version), 2 (n.leader), 0x1b8005f4bba (n.zxid), 0x1 (n.round), LOOKING
> (n.state), 2 (n.sid), 0x1b8 (n.peerEPoch), LOOKING (my state)0 (n.config
> version)
> 2020-03-11 20:23:35,742 [myid:2] - WARN
> [SendWorker:3:QuorumCnxManager$SendWorker@1143] - Interrupted while waiting
> for message on queue
> java.lang.InterruptedException
> at
> java.util.concurrent.locks.AbstractQueuedSynchronizer$ConditionObject.reportInterruptAfterWait(AbstractQueuedSynchronizer.java:2014)
> at
> java.util.concurrent.locks.AbstractQueuedSynchronizer$ConditionObject.awaitNanos(AbstractQueuedSynchronizer.java:2088)
> at
> java.util.concurrent.ArrayBlockingQueue.poll(ArrayBlockingQueue.java:418)
> at
> org.apache.zookeeper.server.quorum.QuorumCnxManager.pollSendQueue(QuorumCnxManager.java:1294)
> at
> org.apache.zookeeper.server.quorum.QuorumCnxManager.access$700(QuorumCnxManager.java:82)
> at
> org.apache.zookeeper.server.quorum.QuorumCnxManager$SendWorker.run(QuorumCnxManager.java:1131)
> 2020-03-11 20:23:35,744 [myid:2] - WARN
> [SendWorker:3:QuorumCnxManager$SendWorker@1153] - Send worker leaving thread
> id 3 my id = 2
> 2020-03-11 20:23:35,745 [myid:2] - WARN
> [RecvWorker:3:QuorumCnxManager$RecvWorker@1230] - Interrupting
> SendWorker{code}
> The
[jira] [Commented] (ZOOKEEPER-3756) Members failing to rejoin quorum
[
https://issues.apache.org/jira/browse/ZOOKEEPER-3756?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17065026#comment-17065026
]
Norbert Kalmár commented on ZOOKEEPER-3756:
---
[~lasaro] are you sure it's the same problem? Why does node3 get elected as
Leader after a restart, when node2 already taken over? That seems strange to
me. A log would be helpful.
I would suggest you write to u...@zookeeper.apache.org with more details and
logs.
[~symat] did the fix, perhaps he can also chip in, as he has more insight on
this one.
> Members failing to rejoin quorum
>
>
> Key: ZOOKEEPER-3756
> URL: https://issues.apache.org/jira/browse/ZOOKEEPER-3756
> Project: ZooKeeper
> Issue Type: Improvement
> Components: leaderElection
>Affects Versions: 3.5.6, 3.5.7
>Reporter: Dai Shi
>Assignee: Mate Szalay-Beko
>Priority: Major
> Labels: pull-request-available
> Fix For: 3.6.1
>
> Attachments: Dockerfile, configmap.yaml, docker-entrypoint.sh,
> jmx.yaml, zoo-0.log, zoo-1.log, zoo-2.log, zoo-service.yaml, zookeeper.yaml
>
> Time Spent: 3h
> Remaining Estimate: 0h
>
> Not sure if this is the place to ask, please close if it's not.
> I am seeing some behavior that I can't explain since upgrading to 3.5:
> In a 5 member quorum, when server 3 is the leader and each server has this in
> their configuration:
> {code:java}
> server.1=100.71.255.254:2888:3888:participant;2181
> server.2=100.71.255.253:2888:3888:participant;2181
> server.3=100.71.255.252:2888:3888:participant;2181
> server.4=100.71.255.251:2888:3888:participant;2181
> server.5=100.71.255.250:2888:3888:participant;2181{code}
> If servers 1 or 2 are restarted, they fail to rejoin the quorum with this in
> the logs:
> {code:java}
> 2020-03-11 20:23:35,720 [myid:2] - INFO
> [QuorumPeer[myid=2](plain=0.0.0.0:2181)(secure=disabled):QuorumPeer@1175] -
> LOOKING
> 2020-03-11 20:23:35,721 [myid:2] - INFO
> [QuorumPeer[myid=2](plain=0.0.0.0:2181)(secure=disabled):FastLeaderElection@885]
> - New election. My id = 2, proposed zxid=0x1b8005f4bba
> 2020-03-11 20:23:35,733 [myid:2] - INFO
> [WorkerSender[myid=2]:QuorumCnxManager@438] - Have smaller server identifier,
> so dropping the connection: (3, 2)
> 2020-03-11 20:23:35,734 [myid:2] - INFO
> [0.0.0.0/0.0.0.0:3888:QuorumCnxManager$Listener@924] - Received connection
> request 100.126.116.201:36140
> 2020-03-11 20:23:35,735 [myid:2] - INFO
> [WorkerSender[myid=2]:QuorumCnxManager@438] - Have smaller server identifier,
> so dropping the connection: (4, 2)
> 2020-03-11 20:23:35,740 [myid:2] - INFO
> [WorkerSender[myid=2]:QuorumCnxManager@438] - Have smaller server identifier,
> so dropping the connection: (5, 2)
> 2020-03-11 20:23:35,740 [myid:2] - INFO
> [0.0.0.0/0.0.0.0:3888:QuorumCnxManager$Listener@924] - Received connection
> request 100.126.116.201:36142
> 2020-03-11 20:23:35,740 [myid:2] - INFO
> [WorkerReceiver[myid=2]:FastLeaderElection@679] - Notification: 2 (message
> format version), 2 (n.leader), 0x1b8005f4bba (n.zxid), 0x1 (n.round), LOOKING
> (n.state), 2 (n.sid), 0x1b8 (n.peerEPoch), LOOKING (my state)0 (n.config
> version)
> 2020-03-11 20:23:35,742 [myid:2] - WARN
> [SendWorker:3:QuorumCnxManager$SendWorker@1143] - Interrupted while waiting
> for message on queue
> java.lang.InterruptedException
> at
> java.util.concurrent.locks.AbstractQueuedSynchronizer$ConditionObject.reportInterruptAfterWait(AbstractQueuedSynchronizer.java:2014)
> at
> java.util.concurrent.locks.AbstractQueuedSynchronizer$ConditionObject.awaitNanos(AbstractQueuedSynchronizer.java:2088)
> at
> java.util.concurrent.ArrayBlockingQueue.poll(ArrayBlockingQueue.java:418)
> at
> org.apache.zookeeper.server.quorum.QuorumCnxManager.pollSendQueue(QuorumCnxManager.java:1294)
> at
> org.apache.zookeeper.server.quorum.QuorumCnxManager.access$700(QuorumCnxManager.java:82)
> at
> org.apache.zookeeper.server.quorum.QuorumCnxManager$SendWorker.run(QuorumCnxManager.java:1131)
> 2020-03-11 20:23:35,744 [myid:2] - WARN
> [SendWorker:3:QuorumCnxManager$SendWorker@1153] - Send worker leaving thread
> id 3 my id = 2
> 2020-03-11 20:23:35,745 [myid:2] - WARN
> [RecvWorker:3:QuorumCnxManager$RecvWorker@1230] - Interrupting
> SendWorker{code}
> The only way I can seem to get them to rejoin the quorum is to restart the
> leader.
> However, if I remove server 4 and 5 from the configuration of server 1 or 2
> (so only servers 1, 2, and 3 remain in the configuration file), then they can
> rejoin the quorum fine. Is this expected and am I doing something wrong? Any
> help or explanation would be greatly appreciated. Thank you.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
[jira] [Commented] (ZOOKEEPER-3756) Members failing to rejoin quorum
[
https://issues.apache.org/jira/browse/ZOOKEEPER-3756?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17065017#comment-17065017
]
Lasaro Camargos commented on ZOOKEEPER-3756:
Dear all,
currently, I am consistently facing the following scenario while running 3.5.5
and 3.5.7, which I believe is related to this bug:
3 nodes up.
Node 3 stop -> node 2 is elected; node 1 follows.
Node 3 start -> node 3 elected the leader; node 2 follows; node 1 is unable to
elect.
Node 1 stop and start -> node 1 rejoins the quorum.
Node 2 stop and start -> node 2 is unable to elect.
Node 1 stop and start -> node 2 joins the quorum; node 1 joins the quorum
Node 2 stop and start -> node 2 unable to join the quorum
Node 3 stop and start -> node 3 elected the leader; node 2 follows; node 1 is
unable to elect.
Reducing the cnxTimeout value didn't change the behavior.
I tested with this fix and now it is now worse; after a round of restarts,
there doesn't seem to have anything I can to make node 1 join finish the
election.
> Members failing to rejoin quorum
>
>
> Key: ZOOKEEPER-3756
> URL: https://issues.apache.org/jira/browse/ZOOKEEPER-3756
> Project: ZooKeeper
> Issue Type: Improvement
> Components: leaderElection
>Affects Versions: 3.5.6, 3.5.7
>Reporter: Dai Shi
>Assignee: Mate Szalay-Beko
>Priority: Major
> Labels: pull-request-available
> Fix For: 3.6.1
>
> Attachments: Dockerfile, configmap.yaml, docker-entrypoint.sh,
> jmx.yaml, zoo-0.log, zoo-1.log, zoo-2.log, zoo-service.yaml, zookeeper.yaml
>
> Time Spent: 3h
> Remaining Estimate: 0h
>
> Not sure if this is the place to ask, please close if it's not.
> I am seeing some behavior that I can't explain since upgrading to 3.5:
> In a 5 member quorum, when server 3 is the leader and each server has this in
> their configuration:
> {code:java}
> server.1=100.71.255.254:2888:3888:participant;2181
> server.2=100.71.255.253:2888:3888:participant;2181
> server.3=100.71.255.252:2888:3888:participant;2181
> server.4=100.71.255.251:2888:3888:participant;2181
> server.5=100.71.255.250:2888:3888:participant;2181{code}
> If servers 1 or 2 are restarted, they fail to rejoin the quorum with this in
> the logs:
> {code:java}
> 2020-03-11 20:23:35,720 [myid:2] - INFO
> [QuorumPeer[myid=2](plain=0.0.0.0:2181)(secure=disabled):QuorumPeer@1175] -
> LOOKING
> 2020-03-11 20:23:35,721 [myid:2] - INFO
> [QuorumPeer[myid=2](plain=0.0.0.0:2181)(secure=disabled):FastLeaderElection@885]
> - New election. My id = 2, proposed zxid=0x1b8005f4bba
> 2020-03-11 20:23:35,733 [myid:2] - INFO
> [WorkerSender[myid=2]:QuorumCnxManager@438] - Have smaller server identifier,
> so dropping the connection: (3, 2)
> 2020-03-11 20:23:35,734 [myid:2] - INFO
> [0.0.0.0/0.0.0.0:3888:QuorumCnxManager$Listener@924] - Received connection
> request 100.126.116.201:36140
> 2020-03-11 20:23:35,735 [myid:2] - INFO
> [WorkerSender[myid=2]:QuorumCnxManager@438] - Have smaller server identifier,
> so dropping the connection: (4, 2)
> 2020-03-11 20:23:35,740 [myid:2] - INFO
> [WorkerSender[myid=2]:QuorumCnxManager@438] - Have smaller server identifier,
> so dropping the connection: (5, 2)
> 2020-03-11 20:23:35,740 [myid:2] - INFO
> [0.0.0.0/0.0.0.0:3888:QuorumCnxManager$Listener@924] - Received connection
> request 100.126.116.201:36142
> 2020-03-11 20:23:35,740 [myid:2] - INFO
> [WorkerReceiver[myid=2]:FastLeaderElection@679] - Notification: 2 (message
> format version), 2 (n.leader), 0x1b8005f4bba (n.zxid), 0x1 (n.round), LOOKING
> (n.state), 2 (n.sid), 0x1b8 (n.peerEPoch), LOOKING (my state)0 (n.config
> version)
> 2020-03-11 20:23:35,742 [myid:2] - WARN
> [SendWorker:3:QuorumCnxManager$SendWorker@1143] - Interrupted while waiting
> for message on queue
> java.lang.InterruptedException
> at
> java.util.concurrent.locks.AbstractQueuedSynchronizer$ConditionObject.reportInterruptAfterWait(AbstractQueuedSynchronizer.java:2014)
> at
> java.util.concurrent.locks.AbstractQueuedSynchronizer$ConditionObject.awaitNanos(AbstractQueuedSynchronizer.java:2088)
> at
> java.util.concurrent.ArrayBlockingQueue.poll(ArrayBlockingQueue.java:418)
> at
> org.apache.zookeeper.server.quorum.QuorumCnxManager.pollSendQueue(QuorumCnxManager.java:1294)
> at
> org.apache.zookeeper.server.quorum.QuorumCnxManager.access$700(QuorumCnxManager.java:82)
> at
> org.apache.zookeeper.server.quorum.QuorumCnxManager$SendWorker.run(QuorumCnxManager.java:1131)
> 2020-03-11 20:23:35,744 [myid:2] - WARN
> [SendWorker:3:QuorumCnxManager$SendWorker@1153] - Send worker leaving thread
> id 3 my id = 2
> 2020-03-11 20:23:35,745 [myid:2] - WARN
> [RecvWorker:3:QuorumCnxManager$RecvWorker@1230] - Interrupting
>
[jira] [Commented] (ZOOKEEPER-3756) Members failing to rejoin quorum
[
https://issues.apache.org/jira/browse/ZOOKEEPER-3756?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17062900#comment-17062900
]
Dai Shi commented on ZOOKEEPER-3756:
OK, thank you!
> Members failing to rejoin quorum
>
>
> Key: ZOOKEEPER-3756
> URL: https://issues.apache.org/jira/browse/ZOOKEEPER-3756
> Project: ZooKeeper
> Issue Type: Improvement
> Components: leaderElection
>Affects Versions: 3.5.6, 3.5.7
>Reporter: Dai Shi
>Assignee: Mate Szalay-Beko
>Priority: Major
> Labels: pull-request-available
> Fix For: 3.6.1, 3.5.8
>
> Attachments: Dockerfile, configmap.yaml, docker-entrypoint.sh,
> jmx.yaml, zoo-0.log, zoo-1.log, zoo-2.log, zoo-service.yaml, zookeeper.yaml
>
> Time Spent: 1.5h
> Remaining Estimate: 0h
>
> Not sure if this is the place to ask, please close if it's not.
> I am seeing some behavior that I can't explain since upgrading to 3.5:
> In a 5 member quorum, when server 3 is the leader and each server has this in
> their configuration:
> {code:java}
> server.1=100.71.255.254:2888:3888:participant;2181
> server.2=100.71.255.253:2888:3888:participant;2181
> server.3=100.71.255.252:2888:3888:participant;2181
> server.4=100.71.255.251:2888:3888:participant;2181
> server.5=100.71.255.250:2888:3888:participant;2181{code}
> If servers 1 or 2 are restarted, they fail to rejoin the quorum with this in
> the logs:
> {code:java}
> 2020-03-11 20:23:35,720 [myid:2] - INFO
> [QuorumPeer[myid=2](plain=0.0.0.0:2181)(secure=disabled):QuorumPeer@1175] -
> LOOKING
> 2020-03-11 20:23:35,721 [myid:2] - INFO
> [QuorumPeer[myid=2](plain=0.0.0.0:2181)(secure=disabled):FastLeaderElection@885]
> - New election. My id = 2, proposed zxid=0x1b8005f4bba
> 2020-03-11 20:23:35,733 [myid:2] - INFO
> [WorkerSender[myid=2]:QuorumCnxManager@438] - Have smaller server identifier,
> so dropping the connection: (3, 2)
> 2020-03-11 20:23:35,734 [myid:2] - INFO
> [0.0.0.0/0.0.0.0:3888:QuorumCnxManager$Listener@924] - Received connection
> request 100.126.116.201:36140
> 2020-03-11 20:23:35,735 [myid:2] - INFO
> [WorkerSender[myid=2]:QuorumCnxManager@438] - Have smaller server identifier,
> so dropping the connection: (4, 2)
> 2020-03-11 20:23:35,740 [myid:2] - INFO
> [WorkerSender[myid=2]:QuorumCnxManager@438] - Have smaller server identifier,
> so dropping the connection: (5, 2)
> 2020-03-11 20:23:35,740 [myid:2] - INFO
> [0.0.0.0/0.0.0.0:3888:QuorumCnxManager$Listener@924] - Received connection
> request 100.126.116.201:36142
> 2020-03-11 20:23:35,740 [myid:2] - INFO
> [WorkerReceiver[myid=2]:FastLeaderElection@679] - Notification: 2 (message
> format version), 2 (n.leader), 0x1b8005f4bba (n.zxid), 0x1 (n.round), LOOKING
> (n.state), 2 (n.sid), 0x1b8 (n.peerEPoch), LOOKING (my state)0 (n.config
> version)
> 2020-03-11 20:23:35,742 [myid:2] - WARN
> [SendWorker:3:QuorumCnxManager$SendWorker@1143] - Interrupted while waiting
> for message on queue
> java.lang.InterruptedException
> at
> java.util.concurrent.locks.AbstractQueuedSynchronizer$ConditionObject.reportInterruptAfterWait(AbstractQueuedSynchronizer.java:2014)
> at
> java.util.concurrent.locks.AbstractQueuedSynchronizer$ConditionObject.awaitNanos(AbstractQueuedSynchronizer.java:2088)
> at
> java.util.concurrent.ArrayBlockingQueue.poll(ArrayBlockingQueue.java:418)
> at
> org.apache.zookeeper.server.quorum.QuorumCnxManager.pollSendQueue(QuorumCnxManager.java:1294)
> at
> org.apache.zookeeper.server.quorum.QuorumCnxManager.access$700(QuorumCnxManager.java:82)
> at
> org.apache.zookeeper.server.quorum.QuorumCnxManager$SendWorker.run(QuorumCnxManager.java:1131)
> 2020-03-11 20:23:35,744 [myid:2] - WARN
> [SendWorker:3:QuorumCnxManager$SendWorker@1153] - Send worker leaving thread
> id 3 my id = 2
> 2020-03-11 20:23:35,745 [myid:2] - WARN
> [RecvWorker:3:QuorumCnxManager$RecvWorker@1230] - Interrupting
> SendWorker{code}
> The only way I can seem to get them to rejoin the quorum is to restart the
> leader.
> However, if I remove server 4 and 5 from the configuration of server 1 or 2
> (so only servers 1, 2, and 3 remain in the configuration file), then they can
> rejoin the quorum fine. Is this expected and am I doing something wrong? Any
> help or explanation would be greatly appreciated. Thank you.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
[jira] [Commented] (ZOOKEEPER-3756) Members failing to rejoin quorum
[
https://issues.apache.org/jira/browse/ZOOKEEPER-3756?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17062621#comment-17062621
]
Mate Szalay-Beko commented on ZOOKEEPER-3756:
-
I managed to reproduce the same behaviour / log messages you saw with some unit
tests, so I consider this to be enough for reproduction / testing. I submitted
a PR for the master & 3.6 branches. Once it got reviewed, I will also submit
the fix on the branch 3.5.
> Members failing to rejoin quorum
>
>
> Key: ZOOKEEPER-3756
> URL: https://issues.apache.org/jira/browse/ZOOKEEPER-3756
> Project: ZooKeeper
> Issue Type: Improvement
> Components: leaderElection
>Affects Versions: 3.5.6, 3.5.7
>Reporter: Dai Shi
>Assignee: Mate Szalay-Beko
>Priority: Major
> Labels: pull-request-available
> Attachments: Dockerfile, configmap.yaml, docker-entrypoint.sh,
> jmx.yaml, zoo-0.log, zoo-1.log, zoo-2.log, zoo-service.yaml, zookeeper.yaml
>
> Time Spent: 10m
> Remaining Estimate: 0h
>
> Not sure if this is the place to ask, please close if it's not.
> I am seeing some behavior that I can't explain since upgrading to 3.5:
> In a 5 member quorum, when server 3 is the leader and each server has this in
> their configuration:
> {code:java}
> server.1=100.71.255.254:2888:3888:participant;2181
> server.2=100.71.255.253:2888:3888:participant;2181
> server.3=100.71.255.252:2888:3888:participant;2181
> server.4=100.71.255.251:2888:3888:participant;2181
> server.5=100.71.255.250:2888:3888:participant;2181{code}
> If servers 1 or 2 are restarted, they fail to rejoin the quorum with this in
> the logs:
> {code:java}
> 2020-03-11 20:23:35,720 [myid:2] - INFO
> [QuorumPeer[myid=2](plain=0.0.0.0:2181)(secure=disabled):QuorumPeer@1175] -
> LOOKING
> 2020-03-11 20:23:35,721 [myid:2] - INFO
> [QuorumPeer[myid=2](plain=0.0.0.0:2181)(secure=disabled):FastLeaderElection@885]
> - New election. My id = 2, proposed zxid=0x1b8005f4bba
> 2020-03-11 20:23:35,733 [myid:2] - INFO
> [WorkerSender[myid=2]:QuorumCnxManager@438] - Have smaller server identifier,
> so dropping the connection: (3, 2)
> 2020-03-11 20:23:35,734 [myid:2] - INFO
> [0.0.0.0/0.0.0.0:3888:QuorumCnxManager$Listener@924] - Received connection
> request 100.126.116.201:36140
> 2020-03-11 20:23:35,735 [myid:2] - INFO
> [WorkerSender[myid=2]:QuorumCnxManager@438] - Have smaller server identifier,
> so dropping the connection: (4, 2)
> 2020-03-11 20:23:35,740 [myid:2] - INFO
> [WorkerSender[myid=2]:QuorumCnxManager@438] - Have smaller server identifier,
> so dropping the connection: (5, 2)
> 2020-03-11 20:23:35,740 [myid:2] - INFO
> [0.0.0.0/0.0.0.0:3888:QuorumCnxManager$Listener@924] - Received connection
> request 100.126.116.201:36142
> 2020-03-11 20:23:35,740 [myid:2] - INFO
> [WorkerReceiver[myid=2]:FastLeaderElection@679] - Notification: 2 (message
> format version), 2 (n.leader), 0x1b8005f4bba (n.zxid), 0x1 (n.round), LOOKING
> (n.state), 2 (n.sid), 0x1b8 (n.peerEPoch), LOOKING (my state)0 (n.config
> version)
> 2020-03-11 20:23:35,742 [myid:2] - WARN
> [SendWorker:3:QuorumCnxManager$SendWorker@1143] - Interrupted while waiting
> for message on queue
> java.lang.InterruptedException
> at
> java.util.concurrent.locks.AbstractQueuedSynchronizer$ConditionObject.reportInterruptAfterWait(AbstractQueuedSynchronizer.java:2014)
> at
> java.util.concurrent.locks.AbstractQueuedSynchronizer$ConditionObject.awaitNanos(AbstractQueuedSynchronizer.java:2088)
> at
> java.util.concurrent.ArrayBlockingQueue.poll(ArrayBlockingQueue.java:418)
> at
> org.apache.zookeeper.server.quorum.QuorumCnxManager.pollSendQueue(QuorumCnxManager.java:1294)
> at
> org.apache.zookeeper.server.quorum.QuorumCnxManager.access$700(QuorumCnxManager.java:82)
> at
> org.apache.zookeeper.server.quorum.QuorumCnxManager$SendWorker.run(QuorumCnxManager.java:1131)
> 2020-03-11 20:23:35,744 [myid:2] - WARN
> [SendWorker:3:QuorumCnxManager$SendWorker@1153] - Send worker leaving thread
> id 3 my id = 2
> 2020-03-11 20:23:35,745 [myid:2] - WARN
> [RecvWorker:3:QuorumCnxManager$RecvWorker@1230] - Interrupting
> SendWorker{code}
> The only way I can seem to get them to rejoin the quorum is to restart the
> leader.
> However, if I remove server 4 and 5 from the configuration of server 1 or 2
> (so only servers 1, 2, and 3 remain in the configuration file), then they can
> rejoin the quorum fine. Is this expected and am I doing something wrong? Any
> help or explanation would be greatly appreciated. Thank you.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
[jira] [Commented] (ZOOKEEPER-3756) Members failing to rejoin quorum
[
https://issues.apache.org/jira/browse/ZOOKEEPER-3756?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17062083#comment-17062083
]
Dai Shi commented on ZOOKEEPER-3756:
OK thanks, let me know if you need any clarifications on the config files I
sent.
> Members failing to rejoin quorum
>
>
> Key: ZOOKEEPER-3756
> URL: https://issues.apache.org/jira/browse/ZOOKEEPER-3756
> Project: ZooKeeper
> Issue Type: Improvement
> Components: leaderElection
>Affects Versions: 3.5.6, 3.5.7
>Reporter: Dai Shi
>Assignee: Mate Szalay-Beko
>Priority: Major
> Attachments: Dockerfile, configmap.yaml, docker-entrypoint.sh,
> jmx.yaml, zoo-0.log, zoo-1.log, zoo-2.log, zoo-service.yaml, zookeeper.yaml
>
>
> Not sure if this is the place to ask, please close if it's not.
> I am seeing some behavior that I can't explain since upgrading to 3.5:
> In a 5 member quorum, when server 3 is the leader and each server has this in
> their configuration:
> {code:java}
> server.1=100.71.255.254:2888:3888:participant;2181
> server.2=100.71.255.253:2888:3888:participant;2181
> server.3=100.71.255.252:2888:3888:participant;2181
> server.4=100.71.255.251:2888:3888:participant;2181
> server.5=100.71.255.250:2888:3888:participant;2181{code}
> If servers 1 or 2 are restarted, they fail to rejoin the quorum with this in
> the logs:
> {code:java}
> 2020-03-11 20:23:35,720 [myid:2] - INFO
> [QuorumPeer[myid=2](plain=0.0.0.0:2181)(secure=disabled):QuorumPeer@1175] -
> LOOKING
> 2020-03-11 20:23:35,721 [myid:2] - INFO
> [QuorumPeer[myid=2](plain=0.0.0.0:2181)(secure=disabled):FastLeaderElection@885]
> - New election. My id = 2, proposed zxid=0x1b8005f4bba
> 2020-03-11 20:23:35,733 [myid:2] - INFO
> [WorkerSender[myid=2]:QuorumCnxManager@438] - Have smaller server identifier,
> so dropping the connection: (3, 2)
> 2020-03-11 20:23:35,734 [myid:2] - INFO
> [0.0.0.0/0.0.0.0:3888:QuorumCnxManager$Listener@924] - Received connection
> request 100.126.116.201:36140
> 2020-03-11 20:23:35,735 [myid:2] - INFO
> [WorkerSender[myid=2]:QuorumCnxManager@438] - Have smaller server identifier,
> so dropping the connection: (4, 2)
> 2020-03-11 20:23:35,740 [myid:2] - INFO
> [WorkerSender[myid=2]:QuorumCnxManager@438] - Have smaller server identifier,
> so dropping the connection: (5, 2)
> 2020-03-11 20:23:35,740 [myid:2] - INFO
> [0.0.0.0/0.0.0.0:3888:QuorumCnxManager$Listener@924] - Received connection
> request 100.126.116.201:36142
> 2020-03-11 20:23:35,740 [myid:2] - INFO
> [WorkerReceiver[myid=2]:FastLeaderElection@679] - Notification: 2 (message
> format version), 2 (n.leader), 0x1b8005f4bba (n.zxid), 0x1 (n.round), LOOKING
> (n.state), 2 (n.sid), 0x1b8 (n.peerEPoch), LOOKING (my state)0 (n.config
> version)
> 2020-03-11 20:23:35,742 [myid:2] - WARN
> [SendWorker:3:QuorumCnxManager$SendWorker@1143] - Interrupted while waiting
> for message on queue
> java.lang.InterruptedException
> at
> java.util.concurrent.locks.AbstractQueuedSynchronizer$ConditionObject.reportInterruptAfterWait(AbstractQueuedSynchronizer.java:2014)
> at
> java.util.concurrent.locks.AbstractQueuedSynchronizer$ConditionObject.awaitNanos(AbstractQueuedSynchronizer.java:2088)
> at
> java.util.concurrent.ArrayBlockingQueue.poll(ArrayBlockingQueue.java:418)
> at
> org.apache.zookeeper.server.quorum.QuorumCnxManager.pollSendQueue(QuorumCnxManager.java:1294)
> at
> org.apache.zookeeper.server.quorum.QuorumCnxManager.access$700(QuorumCnxManager.java:82)
> at
> org.apache.zookeeper.server.quorum.QuorumCnxManager$SendWorker.run(QuorumCnxManager.java:1131)
> 2020-03-11 20:23:35,744 [myid:2] - WARN
> [SendWorker:3:QuorumCnxManager$SendWorker@1153] - Send worker leaving thread
> id 3 my id = 2
> 2020-03-11 20:23:35,745 [myid:2] - WARN
> [RecvWorker:3:QuorumCnxManager$RecvWorker@1230] - Interrupting
> SendWorker{code}
> The only way I can seem to get them to rejoin the quorum is to restart the
> leader.
> However, if I remove server 4 and 5 from the configuration of server 1 or 2
> (so only servers 1, 2, and 3 remain in the configuration file), then they can
> rejoin the quorum fine. Is this expected and am I doing something wrong? Any
> help or explanation would be greatly appreciated. Thank you.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
[jira] [Commented] (ZOOKEEPER-3756) Members failing to rejoin quorum
[
https://issues.apache.org/jira/browse/ZOOKEEPER-3756?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17061481#comment-17061481
]
Mate Szalay-Beko commented on ZOOKEEPER-3756:
-
I am not familiar with Kubernetes enough to understand all the yaml files you
sent :)
so I started with the plain docker part. I was able to build your dockerfile
(modified a bit, removed the GPG stuff and adding the tar.gz file), then I was
able to start a standalone ZooKeeper Servr container by:
{code:java}
docker build -t zookeeper-3756 .
docker run --rm zookeeper-3756:latest
{code}
the modified Dockerfile:
{code:java}
FROM ubuntu:16.04
# install jre
RUN apt-get update -y && \
apt-get upgrade -y && \
apt-get install -y default-jre gosu netcat-openbsd wget
ARG DISTRO_NAME=zookeeper-3.5.8-SNAPSHOT
ARG ARCHIVE_NAME=apache-$DISTRO_NAME-bin
ENV ZOO_USER=zookeeper \
ZOO_CONF_DIR=/conf \
ZOO_DATA_DIR=/data \
ZOO_DATA_LOG_DIR=/datalog \
ZOO_PORT=2181 \
ZOO_TICK_TIME=2000 \
ZOO_INIT_LIMIT=5 \
ZOO_SYNC_LIMIT=2 \
ZOO_AUTOPURGE_RETAIN_COUNT=50 \
ZOO_AUTOPURGE_INTERVAL=6 \
ZOO_LOG_DIR=/logs \
JMX_CONF_DIR=/etc/jmx
COPY apache-zookeeper-3.5.8-SNAPSHOT-bin.tar.gz /
# Add a user and make dirs
RUN set -x \
&& useradd "$ZOO_USER" \
&& mkdir -p "$ZOO_DATA_LOG_DIR" "$ZOO_DATA_DIR" "$ZOO_CONF_DIR"
"$ZOO_LOG_DIR" "$JMX_CONF_DIR" \
&& chown "$ZOO_USER:$ZOO_USER" "$ZOO_DATA_LOG_DIR" "$ZOO_DATA_DIR"
"$ZOO_CONF_DIR" "$ZOO_LOG_DIR"
# Download Apache Zookeeper, verify its PGP signature, untar and clean up
RUN set -x && \
cd / && \
tar -xzf "$ARCHIVE_NAME.tar.gz" && \
mv "$ARCHIVE_NAME/conf/"* "$ZOO_CONF_DIR" && \
rm "$ARCHIVE_NAME.tar.gz" && \
cd /$ARCHIVE_NAME && \
wget -q
"https://repo1.maven.org/maven2/io/prometheus/jmx/jmx_prometheus_javaagent/0.11.0/jmx_prometheus_javaagent-0.11.0.jar;
WORKDIR $ARCHIVE_NAME
VOLUME ["$ZOO_DATA_DIR", "$ZOO_DATA_LOG_DIR"]
EXPOSE $ZOO_PORT 2888 3888
ENV PATH=$PATH:/$ARCHIVE_NAME/bin \
ZOOCFGDIR=$ZOO_CONF_DIR
COPY docker-entrypoint.sh /
COPY jmx.yaml /etc/jmx/config.yaml
ENTRYPOINT ["/docker-entrypoint.sh"]
CMD ["zkServer.sh", "start-foreground"]
{code}
The fact that the standalone Zookeeper server started is a good sign :)
But I am not sure why you saw the {{"Could not find or load main class"}}
error.
I will try to make a minimal Kubernetes setup where I can reproduce the problem
with the connection timeout, using the original 3.5.7 version.
> Members failing to rejoin quorum
>
>
> Key: ZOOKEEPER-3756
> URL: https://issues.apache.org/jira/browse/ZOOKEEPER-3756
> Project: ZooKeeper
> Issue Type: Improvement
> Components: leaderElection
>Affects Versions: 3.5.6, 3.5.7
>Reporter: Dai Shi
>Assignee: Mate Szalay-Beko
>Priority: Major
> Attachments: Dockerfile, configmap.yaml, docker-entrypoint.sh,
> jmx.yaml, zoo-0.log, zoo-1.log, zoo-2.log, zoo-service.yaml, zookeeper.yaml
>
>
> Not sure if this is the place to ask, please close if it's not.
> I am seeing some behavior that I can't explain since upgrading to 3.5:
> In a 5 member quorum, when server 3 is the leader and each server has this in
> their configuration:
> {code:java}
> server.1=100.71.255.254:2888:3888:participant;2181
> server.2=100.71.255.253:2888:3888:participant;2181
> server.3=100.71.255.252:2888:3888:participant;2181
> server.4=100.71.255.251:2888:3888:participant;2181
> server.5=100.71.255.250:2888:3888:participant;2181{code}
> If servers 1 or 2 are restarted, they fail to rejoin the quorum with this in
> the logs:
> {code:java}
> 2020-03-11 20:23:35,720 [myid:2] - INFO
> [QuorumPeer[myid=2](plain=0.0.0.0:2181)(secure=disabled):QuorumPeer@1175] -
> LOOKING
> 2020-03-11 20:23:35,721 [myid:2] - INFO
> [QuorumPeer[myid=2](plain=0.0.0.0:2181)(secure=disabled):FastLeaderElection@885]
> - New election. My id = 2, proposed zxid=0x1b8005f4bba
> 2020-03-11 20:23:35,733 [myid:2] - INFO
> [WorkerSender[myid=2]:QuorumCnxManager@438] - Have smaller server identifier,
> so dropping the connection: (3, 2)
> 2020-03-11 20:23:35,734 [myid:2] - INFO
> [0.0.0.0/0.0.0.0:3888:QuorumCnxManager$Listener@924] - Received connection
> request 100.126.116.201:36140
> 2020-03-11 20:23:35,735 [myid:2] - INFO
> [WorkerSender[myid=2]:QuorumCnxManager@438] - Have smaller server identifier,
> so dropping the connection: (4, 2)
> 2020-03-11 20:23:35,740 [myid:2] - INFO
> [WorkerSender[myid=2]:QuorumCnxManager@438] - Have smaller server identifier,
> so dropping the connection: (5, 2)
> 2020-03-11 20:23:35,740 [myid:2] - INFO
> [0.0.0.0/0.0.0.0:3888:QuorumCnxManager$Listener@924] - Received connection
> request 100.126.116.201:36142
> 2020-03-11 20:23:35,740 [myid:2] - INFO
> [WorkerReceiver[myid=2]:FastLeaderElection@679]
[jira] [Commented] (ZOOKEEPER-3756) Members failing to rejoin quorum
[
https://issues.apache.org/jira/browse/ZOOKEEPER-3756?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17061248#comment-17061248
]
Dai Shi commented on ZOOKEEPER-3756:
I'm actually getting this error when trying to start the server:
{code:java}
Error: Could not find or load main class
org.apache.zookeeper.server.quorum.QuorumPeerMain{code}
Not sure if I need to change something with our deployment configs or not. All
I changed in our Docker image build is to use the new tar.gz file I downloaded
instead of downloading from the apache website.
> Members failing to rejoin quorum
>
>
> Key: ZOOKEEPER-3756
> URL: https://issues.apache.org/jira/browse/ZOOKEEPER-3756
> Project: ZooKeeper
> Issue Type: Improvement
> Components: leaderElection
>Affects Versions: 3.5.6, 3.5.7
>Reporter: Dai Shi
>Assignee: Mate Szalay-Beko
>Priority: Major
> Attachments: Dockerfile, configmap.yaml, docker-entrypoint.sh,
> jmx.yaml, zoo-0.log, zoo-1.log, zoo-2.log, zoo-service.yaml, zookeeper.yaml
>
>
> Not sure if this is the place to ask, please close if it's not.
> I am seeing some behavior that I can't explain since upgrading to 3.5:
> In a 5 member quorum, when server 3 is the leader and each server has this in
> their configuration:
> {code:java}
> server.1=100.71.255.254:2888:3888:participant;2181
> server.2=100.71.255.253:2888:3888:participant;2181
> server.3=100.71.255.252:2888:3888:participant;2181
> server.4=100.71.255.251:2888:3888:participant;2181
> server.5=100.71.255.250:2888:3888:participant;2181{code}
> If servers 1 or 2 are restarted, they fail to rejoin the quorum with this in
> the logs:
> {code:java}
> 2020-03-11 20:23:35,720 [myid:2] - INFO
> [QuorumPeer[myid=2](plain=0.0.0.0:2181)(secure=disabled):QuorumPeer@1175] -
> LOOKING
> 2020-03-11 20:23:35,721 [myid:2] - INFO
> [QuorumPeer[myid=2](plain=0.0.0.0:2181)(secure=disabled):FastLeaderElection@885]
> - New election. My id = 2, proposed zxid=0x1b8005f4bba
> 2020-03-11 20:23:35,733 [myid:2] - INFO
> [WorkerSender[myid=2]:QuorumCnxManager@438] - Have smaller server identifier,
> so dropping the connection: (3, 2)
> 2020-03-11 20:23:35,734 [myid:2] - INFO
> [0.0.0.0/0.0.0.0:3888:QuorumCnxManager$Listener@924] - Received connection
> request 100.126.116.201:36140
> 2020-03-11 20:23:35,735 [myid:2] - INFO
> [WorkerSender[myid=2]:QuorumCnxManager@438] - Have smaller server identifier,
> so dropping the connection: (4, 2)
> 2020-03-11 20:23:35,740 [myid:2] - INFO
> [WorkerSender[myid=2]:QuorumCnxManager@438] - Have smaller server identifier,
> so dropping the connection: (5, 2)
> 2020-03-11 20:23:35,740 [myid:2] - INFO
> [0.0.0.0/0.0.0.0:3888:QuorumCnxManager$Listener@924] - Received connection
> request 100.126.116.201:36142
> 2020-03-11 20:23:35,740 [myid:2] - INFO
> [WorkerReceiver[myid=2]:FastLeaderElection@679] - Notification: 2 (message
> format version), 2 (n.leader), 0x1b8005f4bba (n.zxid), 0x1 (n.round), LOOKING
> (n.state), 2 (n.sid), 0x1b8 (n.peerEPoch), LOOKING (my state)0 (n.config
> version)
> 2020-03-11 20:23:35,742 [myid:2] - WARN
> [SendWorker:3:QuorumCnxManager$SendWorker@1143] - Interrupted while waiting
> for message on queue
> java.lang.InterruptedException
> at
> java.util.concurrent.locks.AbstractQueuedSynchronizer$ConditionObject.reportInterruptAfterWait(AbstractQueuedSynchronizer.java:2014)
> at
> java.util.concurrent.locks.AbstractQueuedSynchronizer$ConditionObject.awaitNanos(AbstractQueuedSynchronizer.java:2088)
> at
> java.util.concurrent.ArrayBlockingQueue.poll(ArrayBlockingQueue.java:418)
> at
> org.apache.zookeeper.server.quorum.QuorumCnxManager.pollSendQueue(QuorumCnxManager.java:1294)
> at
> org.apache.zookeeper.server.quorum.QuorumCnxManager.access$700(QuorumCnxManager.java:82)
> at
> org.apache.zookeeper.server.quorum.QuorumCnxManager$SendWorker.run(QuorumCnxManager.java:1131)
> 2020-03-11 20:23:35,744 [myid:2] - WARN
> [SendWorker:3:QuorumCnxManager$SendWorker@1153] - Send worker leaving thread
> id 3 my id = 2
> 2020-03-11 20:23:35,745 [myid:2] - WARN
> [RecvWorker:3:QuorumCnxManager$RecvWorker@1230] - Interrupting
> SendWorker{code}
> The only way I can seem to get them to rejoin the quorum is to restart the
> leader.
> However, if I remove server 4 and 5 from the configuration of server 1 or 2
> (so only servers 1, 2, and 3 remain in the configuration file), then they can
> rejoin the quorum fine. Is this expected and am I doing something wrong? Any
> help or explanation would be greatly appreciated. Thank you.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
[jira] [Commented] (ZOOKEEPER-3756) Members failing to rejoin quorum
[
https://issues.apache.org/jira/browse/ZOOKEEPER-3756?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17061215#comment-17061215
]
Dai Shi commented on ZOOKEEPER-3756:
I can try and test it in a little bit, just have to rejigger our Dockerfile to
use this instead.
> Members failing to rejoin quorum
>
>
> Key: ZOOKEEPER-3756
> URL: https://issues.apache.org/jira/browse/ZOOKEEPER-3756
> Project: ZooKeeper
> Issue Type: Improvement
> Components: leaderElection
>Affects Versions: 3.5.6, 3.5.7
>Reporter: Dai Shi
>Assignee: Mate Szalay-Beko
>Priority: Major
> Attachments: Dockerfile, configmap.yaml, docker-entrypoint.sh,
> jmx.yaml, zoo-0.log, zoo-1.log, zoo-2.log, zoo-service.yaml, zookeeper.yaml
>
>
> Not sure if this is the place to ask, please close if it's not.
> I am seeing some behavior that I can't explain since upgrading to 3.5:
> In a 5 member quorum, when server 3 is the leader and each server has this in
> their configuration:
> {code:java}
> server.1=100.71.255.254:2888:3888:participant;2181
> server.2=100.71.255.253:2888:3888:participant;2181
> server.3=100.71.255.252:2888:3888:participant;2181
> server.4=100.71.255.251:2888:3888:participant;2181
> server.5=100.71.255.250:2888:3888:participant;2181{code}
> If servers 1 or 2 are restarted, they fail to rejoin the quorum with this in
> the logs:
> {code:java}
> 2020-03-11 20:23:35,720 [myid:2] - INFO
> [QuorumPeer[myid=2](plain=0.0.0.0:2181)(secure=disabled):QuorumPeer@1175] -
> LOOKING
> 2020-03-11 20:23:35,721 [myid:2] - INFO
> [QuorumPeer[myid=2](plain=0.0.0.0:2181)(secure=disabled):FastLeaderElection@885]
> - New election. My id = 2, proposed zxid=0x1b8005f4bba
> 2020-03-11 20:23:35,733 [myid:2] - INFO
> [WorkerSender[myid=2]:QuorumCnxManager@438] - Have smaller server identifier,
> so dropping the connection: (3, 2)
> 2020-03-11 20:23:35,734 [myid:2] - INFO
> [0.0.0.0/0.0.0.0:3888:QuorumCnxManager$Listener@924] - Received connection
> request 100.126.116.201:36140
> 2020-03-11 20:23:35,735 [myid:2] - INFO
> [WorkerSender[myid=2]:QuorumCnxManager@438] - Have smaller server identifier,
> so dropping the connection: (4, 2)
> 2020-03-11 20:23:35,740 [myid:2] - INFO
> [WorkerSender[myid=2]:QuorumCnxManager@438] - Have smaller server identifier,
> so dropping the connection: (5, 2)
> 2020-03-11 20:23:35,740 [myid:2] - INFO
> [0.0.0.0/0.0.0.0:3888:QuorumCnxManager$Listener@924] - Received connection
> request 100.126.116.201:36142
> 2020-03-11 20:23:35,740 [myid:2] - INFO
> [WorkerReceiver[myid=2]:FastLeaderElection@679] - Notification: 2 (message
> format version), 2 (n.leader), 0x1b8005f4bba (n.zxid), 0x1 (n.round), LOOKING
> (n.state), 2 (n.sid), 0x1b8 (n.peerEPoch), LOOKING (my state)0 (n.config
> version)
> 2020-03-11 20:23:35,742 [myid:2] - WARN
> [SendWorker:3:QuorumCnxManager$SendWorker@1143] - Interrupted while waiting
> for message on queue
> java.lang.InterruptedException
> at
> java.util.concurrent.locks.AbstractQueuedSynchronizer$ConditionObject.reportInterruptAfterWait(AbstractQueuedSynchronizer.java:2014)
> at
> java.util.concurrent.locks.AbstractQueuedSynchronizer$ConditionObject.awaitNanos(AbstractQueuedSynchronizer.java:2088)
> at
> java.util.concurrent.ArrayBlockingQueue.poll(ArrayBlockingQueue.java:418)
> at
> org.apache.zookeeper.server.quorum.QuorumCnxManager.pollSendQueue(QuorumCnxManager.java:1294)
> at
> org.apache.zookeeper.server.quorum.QuorumCnxManager.access$700(QuorumCnxManager.java:82)
> at
> org.apache.zookeeper.server.quorum.QuorumCnxManager$SendWorker.run(QuorumCnxManager.java:1131)
> 2020-03-11 20:23:35,744 [myid:2] - WARN
> [SendWorker:3:QuorumCnxManager$SendWorker@1153] - Send worker leaving thread
> id 3 my id = 2
> 2020-03-11 20:23:35,745 [myid:2] - WARN
> [RecvWorker:3:QuorumCnxManager$RecvWorker@1230] - Interrupting
> SendWorker{code}
> The only way I can seem to get them to rejoin the quorum is to restart the
> leader.
> However, if I remove server 4 and 5 from the configuration of server 1 or 2
> (so only servers 1, 2, and 3 remain in the configuration file), then they can
> rejoin the quorum fine. Is this expected and am I doing something wrong? Any
> help or explanation would be greatly appreciated. Thank you.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
[jira] [Commented] (ZOOKEEPER-3756) Members failing to rejoin quorum
[
https://issues.apache.org/jira/browse/ZOOKEEPER-3756?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17061180#comment-17061180
]
Mate Szalay-Beko commented on ZOOKEEPER-3756:
-
I created a fix, but I am not 100% sure it would solve the problem. I will try
to come up with some automated test to catch this scenario (injecting timeout
somehow into the connection creation), but in the meanwhile I share with you a
patched version:
[https://drive.google.com/open?id=1qcO-anHIPNn6ipD8nmNAiHmuQD7G4Nvf]
([https://github.com/symat/zookeeper/commit/d4159aca0d2aea0a544810d3d9b352c10ea55bdf])
This is based on the branch 3.5, I compiled it with OpenJDK 8u242 (I did it on
Mac, but it shouldn't be a problem).
How easy would be for you to test this version without using the
{{-Dzookeeper.cnxTimeout=500}} workaround?
If it takes much time, then please wait until I verify the patch on my machine
(either manually or with automated tests)...
> Members failing to rejoin quorum
>
>
> Key: ZOOKEEPER-3756
> URL: https://issues.apache.org/jira/browse/ZOOKEEPER-3756
> Project: ZooKeeper
> Issue Type: Improvement
> Components: leaderElection
>Affects Versions: 3.5.6, 3.5.7
>Reporter: Dai Shi
>Assignee: Mate Szalay-Beko
>Priority: Major
> Attachments: Dockerfile, configmap.yaml, docker-entrypoint.sh,
> jmx.yaml, zoo-0.log, zoo-1.log, zoo-2.log, zoo-service.yaml, zookeeper.yaml
>
>
> Not sure if this is the place to ask, please close if it's not.
> I am seeing some behavior that I can't explain since upgrading to 3.5:
> In a 5 member quorum, when server 3 is the leader and each server has this in
> their configuration:
> {code:java}
> server.1=100.71.255.254:2888:3888:participant;2181
> server.2=100.71.255.253:2888:3888:participant;2181
> server.3=100.71.255.252:2888:3888:participant;2181
> server.4=100.71.255.251:2888:3888:participant;2181
> server.5=100.71.255.250:2888:3888:participant;2181{code}
> If servers 1 or 2 are restarted, they fail to rejoin the quorum with this in
> the logs:
> {code:java}
> 2020-03-11 20:23:35,720 [myid:2] - INFO
> [QuorumPeer[myid=2](plain=0.0.0.0:2181)(secure=disabled):QuorumPeer@1175] -
> LOOKING
> 2020-03-11 20:23:35,721 [myid:2] - INFO
> [QuorumPeer[myid=2](plain=0.0.0.0:2181)(secure=disabled):FastLeaderElection@885]
> - New election. My id = 2, proposed zxid=0x1b8005f4bba
> 2020-03-11 20:23:35,733 [myid:2] - INFO
> [WorkerSender[myid=2]:QuorumCnxManager@438] - Have smaller server identifier,
> so dropping the connection: (3, 2)
> 2020-03-11 20:23:35,734 [myid:2] - INFO
> [0.0.0.0/0.0.0.0:3888:QuorumCnxManager$Listener@924] - Received connection
> request 100.126.116.201:36140
> 2020-03-11 20:23:35,735 [myid:2] - INFO
> [WorkerSender[myid=2]:QuorumCnxManager@438] - Have smaller server identifier,
> so dropping the connection: (4, 2)
> 2020-03-11 20:23:35,740 [myid:2] - INFO
> [WorkerSender[myid=2]:QuorumCnxManager@438] - Have smaller server identifier,
> so dropping the connection: (5, 2)
> 2020-03-11 20:23:35,740 [myid:2] - INFO
> [0.0.0.0/0.0.0.0:3888:QuorumCnxManager$Listener@924] - Received connection
> request 100.126.116.201:36142
> 2020-03-11 20:23:35,740 [myid:2] - INFO
> [WorkerReceiver[myid=2]:FastLeaderElection@679] - Notification: 2 (message
> format version), 2 (n.leader), 0x1b8005f4bba (n.zxid), 0x1 (n.round), LOOKING
> (n.state), 2 (n.sid), 0x1b8 (n.peerEPoch), LOOKING (my state)0 (n.config
> version)
> 2020-03-11 20:23:35,742 [myid:2] - WARN
> [SendWorker:3:QuorumCnxManager$SendWorker@1143] - Interrupted while waiting
> for message on queue
> java.lang.InterruptedException
> at
> java.util.concurrent.locks.AbstractQueuedSynchronizer$ConditionObject.reportInterruptAfterWait(AbstractQueuedSynchronizer.java:2014)
> at
> java.util.concurrent.locks.AbstractQueuedSynchronizer$ConditionObject.awaitNanos(AbstractQueuedSynchronizer.java:2088)
> at
> java.util.concurrent.ArrayBlockingQueue.poll(ArrayBlockingQueue.java:418)
> at
> org.apache.zookeeper.server.quorum.QuorumCnxManager.pollSendQueue(QuorumCnxManager.java:1294)
> at
> org.apache.zookeeper.server.quorum.QuorumCnxManager.access$700(QuorumCnxManager.java:82)
> at
> org.apache.zookeeper.server.quorum.QuorumCnxManager$SendWorker.run(QuorumCnxManager.java:1131)
> 2020-03-11 20:23:35,744 [myid:2] - WARN
> [SendWorker:3:QuorumCnxManager$SendWorker@1153] - Send worker leaving thread
> id 3 my id = 2
> 2020-03-11 20:23:35,745 [myid:2] - WARN
> [RecvWorker:3:QuorumCnxManager$RecvWorker@1230] - Interrupting
> SendWorker{code}
> The only way I can seem to get them to rejoin the quorum is to restart the
> leader.
> However, if I remove server 4 and 5 from the configuration of server 1 or 2
> (so only servers 1, 2, and 3
[jira] [Commented] (ZOOKEEPER-3756) Members failing to rejoin quorum
[
https://issues.apache.org/jira/browse/ZOOKEEPER-3756?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17061072#comment-17061072
]
Dai Shi commented on ZOOKEEPER-3756:
OK, I think I may just stick with using the lower connection timeout on 3.5.7
for now. I've been testing in our dev environment. It is too risky to try 3.6.0
in production for us anyway, as we only just moved to 3.5.X recently due to
stability concerns on the earlier 3.5 releases.
I'm looking forward to testing out your fixes to the election protocols though,
so will keep an eye out on your progress. Thanks so much for all your help!
> Members failing to rejoin quorum
>
>
> Key: ZOOKEEPER-3756
> URL: https://issues.apache.org/jira/browse/ZOOKEEPER-3756
> Project: ZooKeeper
> Issue Type: Improvement
> Components: leaderElection
>Affects Versions: 3.5.6, 3.5.7
>Reporter: Dai Shi
>Assignee: Mate Szalay-Beko
>Priority: Major
> Attachments: Dockerfile, configmap.yaml, docker-entrypoint.sh,
> jmx.yaml, zoo-0.log, zoo-1.log, zoo-2.log, zoo-service.yaml, zookeeper.yaml
>
>
> Not sure if this is the place to ask, please close if it's not.
> I am seeing some behavior that I can't explain since upgrading to 3.5:
> In a 5 member quorum, when server 3 is the leader and each server has this in
> their configuration:
> {code:java}
> server.1=100.71.255.254:2888:3888:participant;2181
> server.2=100.71.255.253:2888:3888:participant;2181
> server.3=100.71.255.252:2888:3888:participant;2181
> server.4=100.71.255.251:2888:3888:participant;2181
> server.5=100.71.255.250:2888:3888:participant;2181{code}
> If servers 1 or 2 are restarted, they fail to rejoin the quorum with this in
> the logs:
> {code:java}
> 2020-03-11 20:23:35,720 [myid:2] - INFO
> [QuorumPeer[myid=2](plain=0.0.0.0:2181)(secure=disabled):QuorumPeer@1175] -
> LOOKING
> 2020-03-11 20:23:35,721 [myid:2] - INFO
> [QuorumPeer[myid=2](plain=0.0.0.0:2181)(secure=disabled):FastLeaderElection@885]
> - New election. My id = 2, proposed zxid=0x1b8005f4bba
> 2020-03-11 20:23:35,733 [myid:2] - INFO
> [WorkerSender[myid=2]:QuorumCnxManager@438] - Have smaller server identifier,
> so dropping the connection: (3, 2)
> 2020-03-11 20:23:35,734 [myid:2] - INFO
> [0.0.0.0/0.0.0.0:3888:QuorumCnxManager$Listener@924] - Received connection
> request 100.126.116.201:36140
> 2020-03-11 20:23:35,735 [myid:2] - INFO
> [WorkerSender[myid=2]:QuorumCnxManager@438] - Have smaller server identifier,
> so dropping the connection: (4, 2)
> 2020-03-11 20:23:35,740 [myid:2] - INFO
> [WorkerSender[myid=2]:QuorumCnxManager@438] - Have smaller server identifier,
> so dropping the connection: (5, 2)
> 2020-03-11 20:23:35,740 [myid:2] - INFO
> [0.0.0.0/0.0.0.0:3888:QuorumCnxManager$Listener@924] - Received connection
> request 100.126.116.201:36142
> 2020-03-11 20:23:35,740 [myid:2] - INFO
> [WorkerReceiver[myid=2]:FastLeaderElection@679] - Notification: 2 (message
> format version), 2 (n.leader), 0x1b8005f4bba (n.zxid), 0x1 (n.round), LOOKING
> (n.state), 2 (n.sid), 0x1b8 (n.peerEPoch), LOOKING (my state)0 (n.config
> version)
> 2020-03-11 20:23:35,742 [myid:2] - WARN
> [SendWorker:3:QuorumCnxManager$SendWorker@1143] - Interrupted while waiting
> for message on queue
> java.lang.InterruptedException
> at
> java.util.concurrent.locks.AbstractQueuedSynchronizer$ConditionObject.reportInterruptAfterWait(AbstractQueuedSynchronizer.java:2014)
> at
> java.util.concurrent.locks.AbstractQueuedSynchronizer$ConditionObject.awaitNanos(AbstractQueuedSynchronizer.java:2088)
> at
> java.util.concurrent.ArrayBlockingQueue.poll(ArrayBlockingQueue.java:418)
> at
> org.apache.zookeeper.server.quorum.QuorumCnxManager.pollSendQueue(QuorumCnxManager.java:1294)
> at
> org.apache.zookeeper.server.quorum.QuorumCnxManager.access$700(QuorumCnxManager.java:82)
> at
> org.apache.zookeeper.server.quorum.QuorumCnxManager$SendWorker.run(QuorumCnxManager.java:1131)
> 2020-03-11 20:23:35,744 [myid:2] - WARN
> [SendWorker:3:QuorumCnxManager$SendWorker@1153] - Send worker leaving thread
> id 3 my id = 2
> 2020-03-11 20:23:35,745 [myid:2] - WARN
> [RecvWorker:3:QuorumCnxManager$RecvWorker@1230] - Interrupting
> SendWorker{code}
> The only way I can seem to get them to rejoin the quorum is to restart the
> leader.
> However, if I remove server 4 and 5 from the configuration of server 1 or 2
> (so only servers 1, 2, and 3 remain in the configuration file), then they can
> rejoin the quorum fine. Is this expected and am I doing something wrong? Any
> help or explanation would be greatly appreciated. Thank you.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
[jira] [Commented] (ZOOKEEPER-3756) Members failing to rejoin quorum
[
https://issues.apache.org/jira/browse/ZOOKEEPER-3756?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17060978#comment-17060978
]
Mate Szalay-Beko commented on ZOOKEEPER-3756:
-
{quote}the server IPs in the zookeeper configs are pointing to kubernetes
services, opening a TCP connection to those IPs when there are no backend
endpoints (which is the case when a pod is deleted) will just hang.
{quote}
this explains the problem. And also explains why we don't necessary see this
when using non-kubernetes deployments. I usually test with docker-compose (on a
single host, defining virtual networks), but I haven't seen this issue there
either.
Anyway, I think this is actually a bug in the election protocol implementation,
I should be able to fix this soon in the ZooKeeper code (hopefully in a few
days, when I have a couple of hours to spare).
{quote}Is there anything I should pay attention to while upgrading to 3.6.0?
{quote}
Not much. 3.6.0 should be fully compatible with 3.5. AFAIK even the 3.5 clients
can connect to 3.6.0 server without any problem. The 3.5 server configs should
work with 3.6.0.
{quote}Also is it possible to downgrade back to 3.5.7 afterwards?
{quote}
Yes, until you start to use any 3.6 specific features in the server or in the
clients, you should be able to downgrade anytime. (the snapshot / transaction
log formats are the same) But I actually never tested this, and also I don't
think this would be supported officially.
Although some large companies are already using 3.6 ZooKeeper for some time
(they did use the master branch even before it was released officially), still
using 3.6.0 in production I would consider a bit risky as this is the very
first 3.6 release. But it depends on your use-case / how much you rely on
ZooKeeper / how large traffic to put into it, etc.
But rolling-upgrade or rolling-downgrade with {{multiAddress.enabled=true}}
would not work, as the multiAddress feature will cause backward-incompatible
change in the ZooKeeper internal protocol. If you need continuous availability,
then you have to first upgrade to 3.6.0 with a rolling upgrade, then enable
multiAddress using a rolling restart. And similarly, for downgrade: first
disable multiAddress with a rolling restart, then perform a rolling-downgrade.
Anyway, if {{-Dzookeeper.cnxTimeout=500}} is enough for you, then maybe it
doesn't worth for you to put much more effort into testing the 3.6.0 workaround.
If I do the fix, then it will be part of the next 3.5 and 3.6 releases.
> Members failing to rejoin quorum
>
>
> Key: ZOOKEEPER-3756
> URL: https://issues.apache.org/jira/browse/ZOOKEEPER-3756
> Project: ZooKeeper
> Issue Type: Improvement
> Components: leaderElection
>Affects Versions: 3.5.6, 3.5.7
>Reporter: Dai Shi
>Assignee: Mate Szalay-Beko
>Priority: Major
> Attachments: Dockerfile, configmap.yaml, docker-entrypoint.sh,
> jmx.yaml, zoo-0.log, zoo-1.log, zoo-2.log, zoo-service.yaml, zookeeper.yaml
>
>
> Not sure if this is the place to ask, please close if it's not.
> I am seeing some behavior that I can't explain since upgrading to 3.5:
> In a 5 member quorum, when server 3 is the leader and each server has this in
> their configuration:
> {code:java}
> server.1=100.71.255.254:2888:3888:participant;2181
> server.2=100.71.255.253:2888:3888:participant;2181
> server.3=100.71.255.252:2888:3888:participant;2181
> server.4=100.71.255.251:2888:3888:participant;2181
> server.5=100.71.255.250:2888:3888:participant;2181{code}
> If servers 1 or 2 are restarted, they fail to rejoin the quorum with this in
> the logs:
> {code:java}
> 2020-03-11 20:23:35,720 [myid:2] - INFO
> [QuorumPeer[myid=2](plain=0.0.0.0:2181)(secure=disabled):QuorumPeer@1175] -
> LOOKING
> 2020-03-11 20:23:35,721 [myid:2] - INFO
> [QuorumPeer[myid=2](plain=0.0.0.0:2181)(secure=disabled):FastLeaderElection@885]
> - New election. My id = 2, proposed zxid=0x1b8005f4bba
> 2020-03-11 20:23:35,733 [myid:2] - INFO
> [WorkerSender[myid=2]:QuorumCnxManager@438] - Have smaller server identifier,
> so dropping the connection: (3, 2)
> 2020-03-11 20:23:35,734 [myid:2] - INFO
> [0.0.0.0/0.0.0.0:3888:QuorumCnxManager$Listener@924] - Received connection
> request 100.126.116.201:36140
> 2020-03-11 20:23:35,735 [myid:2] - INFO
> [WorkerSender[myid=2]:QuorumCnxManager@438] - Have smaller server identifier,
> so dropping the connection: (4, 2)
> 2020-03-11 20:23:35,740 [myid:2] - INFO
> [WorkerSender[myid=2]:QuorumCnxManager@438] - Have smaller server identifier,
> so dropping the connection: (5, 2)
> 2020-03-11 20:23:35,740 [myid:2] - INFO
> [0.0.0.0/0.0.0.0:3888:QuorumCnxManager$Listener@924] - Received connection
> request 100.126.116.201:36142
> 2020-03-11 20:23:35,740 [myid:2] - INFO
>
[jira] [Commented] (ZOOKEEPER-3756) Members failing to rejoin quorum
[
https://issues.apache.org/jira/browse/ZOOKEEPER-3756?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17060440#comment-17060440
]
Dai Shi commented on ZOOKEEPER-3756:
I think you are right that kubernetes networking is one of the main issues
here. Because the server IPs in the zookeeper configs are pointing to
kubernetes services, opening a TCP connection to those IPs when there are no
backend endpoints (which is the case when a pod is deleted) will just hang.
I tried running with {{-Dzookeeper.cnxTimeout=500}} and now the cluster stays
down for around 3 to 5 seconds when restarting the leader instead of more than
30 seconds. We may be able to tolerate this duration of downtime as a bandaid.
I can try and build a 3.6.0 docker image and test the multiAddress feature as
well. Is there anything I should pay attention to while upgrading to 3.6.0?
Also is it possible to downgrade back to 3.5.7 afterwards?
> Members failing to rejoin quorum
>
>
> Key: ZOOKEEPER-3756
> URL: https://issues.apache.org/jira/browse/ZOOKEEPER-3756
> Project: ZooKeeper
> Issue Type: Improvement
> Components: leaderElection
>Affects Versions: 3.5.6, 3.5.7
>Reporter: Dai Shi
>Assignee: Mate Szalay-Beko
>Priority: Major
> Attachments: Dockerfile, configmap.yaml, docker-entrypoint.sh,
> jmx.yaml, zoo-0.log, zoo-1.log, zoo-2.log, zoo-service.yaml, zookeeper.yaml
>
>
> Not sure if this is the place to ask, please close if it's not.
> I am seeing some behavior that I can't explain since upgrading to 3.5:
> In a 5 member quorum, when server 3 is the leader and each server has this in
> their configuration:
> {code:java}
> server.1=100.71.255.254:2888:3888:participant;2181
> server.2=100.71.255.253:2888:3888:participant;2181
> server.3=100.71.255.252:2888:3888:participant;2181
> server.4=100.71.255.251:2888:3888:participant;2181
> server.5=100.71.255.250:2888:3888:participant;2181{code}
> If servers 1 or 2 are restarted, they fail to rejoin the quorum with this in
> the logs:
> {code:java}
> 2020-03-11 20:23:35,720 [myid:2] - INFO
> [QuorumPeer[myid=2](plain=0.0.0.0:2181)(secure=disabled):QuorumPeer@1175] -
> LOOKING
> 2020-03-11 20:23:35,721 [myid:2] - INFO
> [QuorumPeer[myid=2](plain=0.0.0.0:2181)(secure=disabled):FastLeaderElection@885]
> - New election. My id = 2, proposed zxid=0x1b8005f4bba
> 2020-03-11 20:23:35,733 [myid:2] - INFO
> [WorkerSender[myid=2]:QuorumCnxManager@438] - Have smaller server identifier,
> so dropping the connection: (3, 2)
> 2020-03-11 20:23:35,734 [myid:2] - INFO
> [0.0.0.0/0.0.0.0:3888:QuorumCnxManager$Listener@924] - Received connection
> request 100.126.116.201:36140
> 2020-03-11 20:23:35,735 [myid:2] - INFO
> [WorkerSender[myid=2]:QuorumCnxManager@438] - Have smaller server identifier,
> so dropping the connection: (4, 2)
> 2020-03-11 20:23:35,740 [myid:2] - INFO
> [WorkerSender[myid=2]:QuorumCnxManager@438] - Have smaller server identifier,
> so dropping the connection: (5, 2)
> 2020-03-11 20:23:35,740 [myid:2] - INFO
> [0.0.0.0/0.0.0.0:3888:QuorumCnxManager$Listener@924] - Received connection
> request 100.126.116.201:36142
> 2020-03-11 20:23:35,740 [myid:2] - INFO
> [WorkerReceiver[myid=2]:FastLeaderElection@679] - Notification: 2 (message
> format version), 2 (n.leader), 0x1b8005f4bba (n.zxid), 0x1 (n.round), LOOKING
> (n.state), 2 (n.sid), 0x1b8 (n.peerEPoch), LOOKING (my state)0 (n.config
> version)
> 2020-03-11 20:23:35,742 [myid:2] - WARN
> [SendWorker:3:QuorumCnxManager$SendWorker@1143] - Interrupted while waiting
> for message on queue
> java.lang.InterruptedException
> at
> java.util.concurrent.locks.AbstractQueuedSynchronizer$ConditionObject.reportInterruptAfterWait(AbstractQueuedSynchronizer.java:2014)
> at
> java.util.concurrent.locks.AbstractQueuedSynchronizer$ConditionObject.awaitNanos(AbstractQueuedSynchronizer.java:2088)
> at
> java.util.concurrent.ArrayBlockingQueue.poll(ArrayBlockingQueue.java:418)
> at
> org.apache.zookeeper.server.quorum.QuorumCnxManager.pollSendQueue(QuorumCnxManager.java:1294)
> at
> org.apache.zookeeper.server.quorum.QuorumCnxManager.access$700(QuorumCnxManager.java:82)
> at
> org.apache.zookeeper.server.quorum.QuorumCnxManager$SendWorker.run(QuorumCnxManager.java:1131)
> 2020-03-11 20:23:35,744 [myid:2] - WARN
> [SendWorker:3:QuorumCnxManager$SendWorker@1153] - Send worker leaving thread
> id 3 my id = 2
> 2020-03-11 20:23:35,745 [myid:2] - WARN
> [RecvWorker:3:QuorumCnxManager$RecvWorker@1230] - Interrupting
> SendWorker{code}
> The only way I can seem to get them to rejoin the quorum is to restart the
> leader.
> However, if I remove server 4 and 5 from the configuration of server 1 or 2
> (so only servers 1, 2, and 3 remain in the configuration
[jira] [Commented] (ZOOKEEPER-3756) Members failing to rejoin quorum
[
https://issues.apache.org/jira/browse/ZOOKEEPER-3756?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17060200#comment-17060200
]
Mate Szalay-Beko commented on ZOOKEEPER-3756:
-
OK, I have a theory... Maybe this is what happens:
- After shutting down the leader, the whole leader election restarts
- ZooKeeper tries to open socket connection to the other ZooKeeper servers by
using synchronized methods, so only one can run a time (see on the master
branch:
https://github.com/apache/zookeeper/blob/a5a4743733b8939464af82c1ee68a593fadbe362/zookeeper-server/src/main/java/org/apache/zookeeper/server/quorum/QuorumCnxManager.java#L688
and
https://github.com/apache/zookeeper/blob/a5a4743733b8939464af82c1ee68a593fadbe362/zookeeper-server/src/main/java/org/apache/zookeeper/server/quorum/QuorumCnxManager.java#L759)
- the default timeout is 5 secs (this is why there is nothing leader election
related log message in your log files for 5 sec, until we hit the timeout of
socket open to server 3)
- by the time when the 5 sec timeout elapsed, the leader election protocol was
also timeouted (but AFAIK it is increasing its internal timeout always? I will
need to verify this)
- after this happens a few time, either the leader election protocol timeout is
increased enough to be able tolerate the 5 sec delay (and/or the fact that the
server-3 restarted and the socket can be opened now) will cause that this block
gets removed and everything goes smoothly after this. But it took 30 seconds,
what is way too long...
The question is, why the socket needs to timeout (wait for 5 sec) and why the
connection doesn't get closed immediately with some 'host unreachable'
exception, what we would expect in case if the server goes down and no IP
connection can be established. Usually we don't see this problem in production,
so I guess it has to do something with Kubernetes networking.
Still, this part needs to be refactored in ZooKeeper, we have to make the
{{connectOne}} asynchronous, what is not an easy task. Actually this is also
something which was suggested in ZOOKEEPER-2164 (but in that ticket there were
other errors fixed in the end).
In the meanwhile there might be some workarounds:
# you can decrease the connection timeout to e.g. 500ms or 1000ms using the
{{-Dzookeeper.cnxTimeout=500'}} system property. I am not sure if it will help,
but I would be glad if you could test it
# an other independent workaround would be using the multiAddress feature of
ZooKeeper 3.6.0, enabling it by {{-Dzookeeper.multiAddress.enabled=true}}. Then
ZooKeeper should periodically check the availability of the currently used
election addresses and kill the socket if the host is unavailable. This way we
might kill the dead socket before the timeout happen. However, it might run
ICMP traffic (ping) in the background, which I am not sure if will be reliable
in kubernetes.
No matter if the workarounds would fix the problem for you or not, I would
suggest to keep this ticket open, and I will try to implement an asynchronous
connection establishment somehow.
> Members failing to rejoin quorum
>
>
> Key: ZOOKEEPER-3756
> URL: https://issues.apache.org/jira/browse/ZOOKEEPER-3756
> Project: ZooKeeper
> Issue Type: Improvement
> Components: leaderElection
>Affects Versions: 3.5.6, 3.5.7
>Reporter: Dai Shi
>Assignee: Mate Szalay-Beko
>Priority: Major
> Attachments: Dockerfile, configmap.yaml, docker-entrypoint.sh,
> jmx.yaml, zoo-0.log, zoo-1.log, zoo-2.log, zoo-service.yaml, zookeeper.yaml
>
>
> Not sure if this is the place to ask, please close if it's not.
> I am seeing some behavior that I can't explain since upgrading to 3.5:
> In a 5 member quorum, when server 3 is the leader and each server has this in
> their configuration:
> {code:java}
> server.1=100.71.255.254:2888:3888:participant;2181
> server.2=100.71.255.253:2888:3888:participant;2181
> server.3=100.71.255.252:2888:3888:participant;2181
> server.4=100.71.255.251:2888:3888:participant;2181
> server.5=100.71.255.250:2888:3888:participant;2181{code}
> If servers 1 or 2 are restarted, they fail to rejoin the quorum with this in
> the logs:
> {code:java}
> 2020-03-11 20:23:35,720 [myid:2] - INFO
> [QuorumPeer[myid=2](plain=0.0.0.0:2181)(secure=disabled):QuorumPeer@1175] -
> LOOKING
> 2020-03-11 20:23:35,721 [myid:2] - INFO
> [QuorumPeer[myid=2](plain=0.0.0.0:2181)(secure=disabled):FastLeaderElection@885]
> - New election. My id = 2, proposed zxid=0x1b8005f4bba
> 2020-03-11 20:23:35,733 [myid:2] - INFO
> [WorkerSender[myid=2]:QuorumCnxManager@438] - Have smaller server identifier,
> so dropping the connection: (3, 2)
> 2020-03-11 20:23:35,734 [myid:2] - INFO
> [0.0.0.0/0.0.0.0:3888:QuorumCnxManager$Listener@924] - Received connection
>
[jira] [Commented] (ZOOKEEPER-3756) Members failing to rejoin quorum
[
https://issues.apache.org/jira/browse/ZOOKEEPER-3756?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17060051#comment-17060051
]
Mate Szalay-Beko commented on ZOOKEEPER-3756:
-
Thanks, it's great that you were able to do this test and sent all the logs. I
need a bit more time to dig into it, I hope I can analyze it deeper and come
back with some answers (possibly questions? :) ) today / tomorrow.
> Members failing to rejoin quorum
>
>
> Key: ZOOKEEPER-3756
> URL: https://issues.apache.org/jira/browse/ZOOKEEPER-3756
> Project: ZooKeeper
> Issue Type: Improvement
> Components: leaderElection
>Affects Versions: 3.5.6, 3.5.7
>Reporter: Dai Shi
>Assignee: Mate Szalay-Beko
>Priority: Major
> Attachments: Dockerfile, configmap.yaml, docker-entrypoint.sh,
> jmx.yaml, zoo-0.log, zoo-1.log, zoo-2.log, zoo-service.yaml, zookeeper.yaml
>
>
> Not sure if this is the place to ask, please close if it's not.
> I am seeing some behavior that I can't explain since upgrading to 3.5:
> In a 5 member quorum, when server 3 is the leader and each server has this in
> their configuration:
> {code:java}
> server.1=100.71.255.254:2888:3888:participant;2181
> server.2=100.71.255.253:2888:3888:participant;2181
> server.3=100.71.255.252:2888:3888:participant;2181
> server.4=100.71.255.251:2888:3888:participant;2181
> server.5=100.71.255.250:2888:3888:participant;2181{code}
> If servers 1 or 2 are restarted, they fail to rejoin the quorum with this in
> the logs:
> {code:java}
> 2020-03-11 20:23:35,720 [myid:2] - INFO
> [QuorumPeer[myid=2](plain=0.0.0.0:2181)(secure=disabled):QuorumPeer@1175] -
> LOOKING
> 2020-03-11 20:23:35,721 [myid:2] - INFO
> [QuorumPeer[myid=2](plain=0.0.0.0:2181)(secure=disabled):FastLeaderElection@885]
> - New election. My id = 2, proposed zxid=0x1b8005f4bba
> 2020-03-11 20:23:35,733 [myid:2] - INFO
> [WorkerSender[myid=2]:QuorumCnxManager@438] - Have smaller server identifier,
> so dropping the connection: (3, 2)
> 2020-03-11 20:23:35,734 [myid:2] - INFO
> [0.0.0.0/0.0.0.0:3888:QuorumCnxManager$Listener@924] - Received connection
> request 100.126.116.201:36140
> 2020-03-11 20:23:35,735 [myid:2] - INFO
> [WorkerSender[myid=2]:QuorumCnxManager@438] - Have smaller server identifier,
> so dropping the connection: (4, 2)
> 2020-03-11 20:23:35,740 [myid:2] - INFO
> [WorkerSender[myid=2]:QuorumCnxManager@438] - Have smaller server identifier,
> so dropping the connection: (5, 2)
> 2020-03-11 20:23:35,740 [myid:2] - INFO
> [0.0.0.0/0.0.0.0:3888:QuorumCnxManager$Listener@924] - Received connection
> request 100.126.116.201:36142
> 2020-03-11 20:23:35,740 [myid:2] - INFO
> [WorkerReceiver[myid=2]:FastLeaderElection@679] - Notification: 2 (message
> format version), 2 (n.leader), 0x1b8005f4bba (n.zxid), 0x1 (n.round), LOOKING
> (n.state), 2 (n.sid), 0x1b8 (n.peerEPoch), LOOKING (my state)0 (n.config
> version)
> 2020-03-11 20:23:35,742 [myid:2] - WARN
> [SendWorker:3:QuorumCnxManager$SendWorker@1143] - Interrupted while waiting
> for message on queue
> java.lang.InterruptedException
> at
> java.util.concurrent.locks.AbstractQueuedSynchronizer$ConditionObject.reportInterruptAfterWait(AbstractQueuedSynchronizer.java:2014)
> at
> java.util.concurrent.locks.AbstractQueuedSynchronizer$ConditionObject.awaitNanos(AbstractQueuedSynchronizer.java:2088)
> at
> java.util.concurrent.ArrayBlockingQueue.poll(ArrayBlockingQueue.java:418)
> at
> org.apache.zookeeper.server.quorum.QuorumCnxManager.pollSendQueue(QuorumCnxManager.java:1294)
> at
> org.apache.zookeeper.server.quorum.QuorumCnxManager.access$700(QuorumCnxManager.java:82)
> at
> org.apache.zookeeper.server.quorum.QuorumCnxManager$SendWorker.run(QuorumCnxManager.java:1131)
> 2020-03-11 20:23:35,744 [myid:2] - WARN
> [SendWorker:3:QuorumCnxManager$SendWorker@1153] - Send worker leaving thread
> id 3 my id = 2
> 2020-03-11 20:23:35,745 [myid:2] - WARN
> [RecvWorker:3:QuorumCnxManager$RecvWorker@1230] - Interrupting
> SendWorker{code}
> The only way I can seem to get them to rejoin the quorum is to restart the
> leader.
> However, if I remove server 4 and 5 from the configuration of server 1 or 2
> (so only servers 1, 2, and 3 remain in the configuration file), then they can
> rejoin the quorum fine. Is this expected and am I doing something wrong? Any
> help or explanation would be greatly appreciated. Thank you.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
[jira] [Commented] (ZOOKEEPER-3756) Members failing to rejoin quorum
[
https://issues.apache.org/jira/browse/ZOOKEEPER-3756?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17059026#comment-17059026
]
Dai Shi commented on ZOOKEEPER-3756:
I repeated the experiment again thought with just 3 members. Here are logs from
each:
[^zoo-0.log]
[^zoo-1.log]
[^zoo-2.log]
Here zoo-2 (ID 3) was the leader the whole time. I did restart zoo-1 just to
get a clean log, so you'll see in zoo-0's logs that zoo-1 was down for a little
bit. Then I restarted zoo-2, which was the leader. I ran this on zoo-0 to
measure the downtime from the client's perspective:
{code:java}
while :; do date; if echo srvr | nc localhost 2181 | grep -q Mode; then true;
else echo DOWN; fi; sleep 0.05; done
{code}
The cluster was from 19:15:34 to 19:16:06. This roughly matches the timestamps
in the logs for these messages:
{code:java}
2020-03-13 19:15:34,945 [myid:2] - WARN
[QuorumPeer[myid=2](plain=0.0.0.0:2181)(secure=disabled):Follower@96] -
Exception when following the leader
{code}
{code:java}
2020-03-13 19:16:05,247 [myid:2] - INFO
[QuorumPeer[myid=2](plain=0.0.0.0:2181)(secure=disabled):Follower@69] -
FOLLOWING - LEADER ELECTION TOOK - 1 MS
{code}
> Members failing to rejoin quorum
>
>
> Key: ZOOKEEPER-3756
> URL: https://issues.apache.org/jira/browse/ZOOKEEPER-3756
> Project: ZooKeeper
> Issue Type: Improvement
> Components: leaderElection
>Affects Versions: 3.5.6, 3.5.7
>Reporter: Dai Shi
>Assignee: Mate Szalay-Beko
>Priority: Major
> Attachments: Dockerfile, configmap.yaml, docker-entrypoint.sh,
> jmx.yaml, zoo-0.log, zoo-1.log, zoo-2.log, zoo-service.yaml, zookeeper.yaml
>
>
> Not sure if this is the place to ask, please close if it's not.
> I am seeing some behavior that I can't explain since upgrading to 3.5:
> In a 5 member quorum, when server 3 is the leader and each server has this in
> their configuration:
> {code:java}
> server.1=100.71.255.254:2888:3888:participant;2181
> server.2=100.71.255.253:2888:3888:participant;2181
> server.3=100.71.255.252:2888:3888:participant;2181
> server.4=100.71.255.251:2888:3888:participant;2181
> server.5=100.71.255.250:2888:3888:participant;2181{code}
> If servers 1 or 2 are restarted, they fail to rejoin the quorum with this in
> the logs:
> {code:java}
> 2020-03-11 20:23:35,720 [myid:2] - INFO
> [QuorumPeer[myid=2](plain=0.0.0.0:2181)(secure=disabled):QuorumPeer@1175] -
> LOOKING
> 2020-03-11 20:23:35,721 [myid:2] - INFO
> [QuorumPeer[myid=2](plain=0.0.0.0:2181)(secure=disabled):FastLeaderElection@885]
> - New election. My id = 2, proposed zxid=0x1b8005f4bba
> 2020-03-11 20:23:35,733 [myid:2] - INFO
> [WorkerSender[myid=2]:QuorumCnxManager@438] - Have smaller server identifier,
> so dropping the connection: (3, 2)
> 2020-03-11 20:23:35,734 [myid:2] - INFO
> [0.0.0.0/0.0.0.0:3888:QuorumCnxManager$Listener@924] - Received connection
> request 100.126.116.201:36140
> 2020-03-11 20:23:35,735 [myid:2] - INFO
> [WorkerSender[myid=2]:QuorumCnxManager@438] - Have smaller server identifier,
> so dropping the connection: (4, 2)
> 2020-03-11 20:23:35,740 [myid:2] - INFO
> [WorkerSender[myid=2]:QuorumCnxManager@438] - Have smaller server identifier,
> so dropping the connection: (5, 2)
> 2020-03-11 20:23:35,740 [myid:2] - INFO
> [0.0.0.0/0.0.0.0:3888:QuorumCnxManager$Listener@924] - Received connection
> request 100.126.116.201:36142
> 2020-03-11 20:23:35,740 [myid:2] - INFO
> [WorkerReceiver[myid=2]:FastLeaderElection@679] - Notification: 2 (message
> format version), 2 (n.leader), 0x1b8005f4bba (n.zxid), 0x1 (n.round), LOOKING
> (n.state), 2 (n.sid), 0x1b8 (n.peerEPoch), LOOKING (my state)0 (n.config
> version)
> 2020-03-11 20:23:35,742 [myid:2] - WARN
> [SendWorker:3:QuorumCnxManager$SendWorker@1143] - Interrupted while waiting
> for message on queue
> java.lang.InterruptedException
> at
> java.util.concurrent.locks.AbstractQueuedSynchronizer$ConditionObject.reportInterruptAfterWait(AbstractQueuedSynchronizer.java:2014)
> at
> java.util.concurrent.locks.AbstractQueuedSynchronizer$ConditionObject.awaitNanos(AbstractQueuedSynchronizer.java:2088)
> at
> java.util.concurrent.ArrayBlockingQueue.poll(ArrayBlockingQueue.java:418)
> at
> org.apache.zookeeper.server.quorum.QuorumCnxManager.pollSendQueue(QuorumCnxManager.java:1294)
> at
> org.apache.zookeeper.server.quorum.QuorumCnxManager.access$700(QuorumCnxManager.java:82)
> at
> org.apache.zookeeper.server.quorum.QuorumCnxManager$SendWorker.run(QuorumCnxManager.java:1131)
> 2020-03-11 20:23:35,744 [myid:2] - WARN
> [SendWorker:3:QuorumCnxManager$SendWorker@1153] - Send worker leaving thread
> id 3 my id = 2
> 2020-03-11 20:23:35,745 [myid:2] - WARN
> [RecvWorker:3:QuorumCnxManager$RecvWorker@1230] - Interrupting
[jira] [Commented] (ZOOKEEPER-3756) Members failing to rejoin quorum
[
https://issues.apache.org/jira/browse/ZOOKEEPER-3756?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17059014#comment-17059014
]
Mate Szalay-Beko commented on ZOOKEEPER-3756:
-
> restarting the leader pod causes the cluster to be down for around 30 seconds
> while it restarts
No, usually a full cluster recovery shouldn't take more than a few seconds. At
least this is what I saw with normal clusters without Kubetnetes.
It would be interesting to see where ZooKeeper spend the 30 seconds. Is the
detection of the loosing of the old leader is slow? Or the new election is this
slow for some reason?
> I'm happy to close this and open a new issue if needed.
I would prefer to keep this one open. Partly to minimize the administration :)
and also there are some discussion / attachment here that is valuable for the
new issue too.
> Members failing to rejoin quorum
>
>
> Key: ZOOKEEPER-3756
> URL: https://issues.apache.org/jira/browse/ZOOKEEPER-3756
> Project: ZooKeeper
> Issue Type: Improvement
> Components: leaderElection
>Affects Versions: 3.5.6, 3.5.7
>Reporter: Dai Shi
>Assignee: Mate Szalay-Beko
>Priority: Major
> Attachments: Dockerfile, configmap.yaml, docker-entrypoint.sh,
> jmx.yaml, zoo-service.yaml, zookeeper.yaml
>
>
> Not sure if this is the place to ask, please close if it's not.
> I am seeing some behavior that I can't explain since upgrading to 3.5:
> In a 5 member quorum, when server 3 is the leader and each server has this in
> their configuration:
> {code:java}
> server.1=100.71.255.254:2888:3888:participant;2181
> server.2=100.71.255.253:2888:3888:participant;2181
> server.3=100.71.255.252:2888:3888:participant;2181
> server.4=100.71.255.251:2888:3888:participant;2181
> server.5=100.71.255.250:2888:3888:participant;2181{code}
> If servers 1 or 2 are restarted, they fail to rejoin the quorum with this in
> the logs:
> {code:java}
> 2020-03-11 20:23:35,720 [myid:2] - INFO
> [QuorumPeer[myid=2](plain=0.0.0.0:2181)(secure=disabled):QuorumPeer@1175] -
> LOOKING
> 2020-03-11 20:23:35,721 [myid:2] - INFO
> [QuorumPeer[myid=2](plain=0.0.0.0:2181)(secure=disabled):FastLeaderElection@885]
> - New election. My id = 2, proposed zxid=0x1b8005f4bba
> 2020-03-11 20:23:35,733 [myid:2] - INFO
> [WorkerSender[myid=2]:QuorumCnxManager@438] - Have smaller server identifier,
> so dropping the connection: (3, 2)
> 2020-03-11 20:23:35,734 [myid:2] - INFO
> [0.0.0.0/0.0.0.0:3888:QuorumCnxManager$Listener@924] - Received connection
> request 100.126.116.201:36140
> 2020-03-11 20:23:35,735 [myid:2] - INFO
> [WorkerSender[myid=2]:QuorumCnxManager@438] - Have smaller server identifier,
> so dropping the connection: (4, 2)
> 2020-03-11 20:23:35,740 [myid:2] - INFO
> [WorkerSender[myid=2]:QuorumCnxManager@438] - Have smaller server identifier,
> so dropping the connection: (5, 2)
> 2020-03-11 20:23:35,740 [myid:2] - INFO
> [0.0.0.0/0.0.0.0:3888:QuorumCnxManager$Listener@924] - Received connection
> request 100.126.116.201:36142
> 2020-03-11 20:23:35,740 [myid:2] - INFO
> [WorkerReceiver[myid=2]:FastLeaderElection@679] - Notification: 2 (message
> format version), 2 (n.leader), 0x1b8005f4bba (n.zxid), 0x1 (n.round), LOOKING
> (n.state), 2 (n.sid), 0x1b8 (n.peerEPoch), LOOKING (my state)0 (n.config
> version)
> 2020-03-11 20:23:35,742 [myid:2] - WARN
> [SendWorker:3:QuorumCnxManager$SendWorker@1143] - Interrupted while waiting
> for message on queue
> java.lang.InterruptedException
> at
> java.util.concurrent.locks.AbstractQueuedSynchronizer$ConditionObject.reportInterruptAfterWait(AbstractQueuedSynchronizer.java:2014)
> at
> java.util.concurrent.locks.AbstractQueuedSynchronizer$ConditionObject.awaitNanos(AbstractQueuedSynchronizer.java:2088)
> at
> java.util.concurrent.ArrayBlockingQueue.poll(ArrayBlockingQueue.java:418)
> at
> org.apache.zookeeper.server.quorum.QuorumCnxManager.pollSendQueue(QuorumCnxManager.java:1294)
> at
> org.apache.zookeeper.server.quorum.QuorumCnxManager.access$700(QuorumCnxManager.java:82)
> at
> org.apache.zookeeper.server.quorum.QuorumCnxManager$SendWorker.run(QuorumCnxManager.java:1131)
> 2020-03-11 20:23:35,744 [myid:2] - WARN
> [SendWorker:3:QuorumCnxManager$SendWorker@1153] - Send worker leaving thread
> id 3 my id = 2
> 2020-03-11 20:23:35,745 [myid:2] - WARN
> [RecvWorker:3:QuorumCnxManager$RecvWorker@1230] - Interrupting
> SendWorker{code}
> The only way I can seem to get them to rejoin the quorum is to restart the
> leader.
> However, if I remove server 4 and 5 from the configuration of server 1 or 2
> (so only servers 1, 2, and 3 remain in the configuration file), then they can
> rejoin the quorum fine. Is this expected and am I doing something wrong? Any
> help or
[jira] [Commented] (ZOOKEEPER-3756) Members failing to rejoin quorum
[
https://issues.apache.org/jira/browse/ZOOKEEPER-3756?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17058988#comment-17058988
]
Mate Szalay-Beko commented on ZOOKEEPER-3756:
-
Thanks for all the details! I will try to reproduce the issue on Monday and
come back with questions / comments soon.
> Members failing to rejoin quorum
>
>
> Key: ZOOKEEPER-3756
> URL: https://issues.apache.org/jira/browse/ZOOKEEPER-3756
> Project: ZooKeeper
> Issue Type: Improvement
> Components: leaderElection
>Affects Versions: 3.5.6, 3.5.7
>Reporter: Dai Shi
>Assignee: Mate Szalay-Beko
>Priority: Major
> Attachments: Dockerfile, configmap.yaml, docker-entrypoint.sh,
> jmx.yaml, zoo-service.yaml, zookeeper.yaml
>
>
> Not sure if this is the place to ask, please close if it's not.
> I am seeing some behavior that I can't explain since upgrading to 3.5:
> In a 5 member quorum, when server 3 is the leader and each server has this in
> their configuration:
> {code:java}
> server.1=100.71.255.254:2888:3888:participant;2181
> server.2=100.71.255.253:2888:3888:participant;2181
> server.3=100.71.255.252:2888:3888:participant;2181
> server.4=100.71.255.251:2888:3888:participant;2181
> server.5=100.71.255.250:2888:3888:participant;2181{code}
> If servers 1 or 2 are restarted, they fail to rejoin the quorum with this in
> the logs:
> {code:java}
> 2020-03-11 20:23:35,720 [myid:2] - INFO
> [QuorumPeer[myid=2](plain=0.0.0.0:2181)(secure=disabled):QuorumPeer@1175] -
> LOOKING
> 2020-03-11 20:23:35,721 [myid:2] - INFO
> [QuorumPeer[myid=2](plain=0.0.0.0:2181)(secure=disabled):FastLeaderElection@885]
> - New election. My id = 2, proposed zxid=0x1b8005f4bba
> 2020-03-11 20:23:35,733 [myid:2] - INFO
> [WorkerSender[myid=2]:QuorumCnxManager@438] - Have smaller server identifier,
> so dropping the connection: (3, 2)
> 2020-03-11 20:23:35,734 [myid:2] - INFO
> [0.0.0.0/0.0.0.0:3888:QuorumCnxManager$Listener@924] - Received connection
> request 100.126.116.201:36140
> 2020-03-11 20:23:35,735 [myid:2] - INFO
> [WorkerSender[myid=2]:QuorumCnxManager@438] - Have smaller server identifier,
> so dropping the connection: (4, 2)
> 2020-03-11 20:23:35,740 [myid:2] - INFO
> [WorkerSender[myid=2]:QuorumCnxManager@438] - Have smaller server identifier,
> so dropping the connection: (5, 2)
> 2020-03-11 20:23:35,740 [myid:2] - INFO
> [0.0.0.0/0.0.0.0:3888:QuorumCnxManager$Listener@924] - Received connection
> request 100.126.116.201:36142
> 2020-03-11 20:23:35,740 [myid:2] - INFO
> [WorkerReceiver[myid=2]:FastLeaderElection@679] - Notification: 2 (message
> format version), 2 (n.leader), 0x1b8005f4bba (n.zxid), 0x1 (n.round), LOOKING
> (n.state), 2 (n.sid), 0x1b8 (n.peerEPoch), LOOKING (my state)0 (n.config
> version)
> 2020-03-11 20:23:35,742 [myid:2] - WARN
> [SendWorker:3:QuorumCnxManager$SendWorker@1143] - Interrupted while waiting
> for message on queue
> java.lang.InterruptedException
> at
> java.util.concurrent.locks.AbstractQueuedSynchronizer$ConditionObject.reportInterruptAfterWait(AbstractQueuedSynchronizer.java:2014)
> at
> java.util.concurrent.locks.AbstractQueuedSynchronizer$ConditionObject.awaitNanos(AbstractQueuedSynchronizer.java:2088)
> at
> java.util.concurrent.ArrayBlockingQueue.poll(ArrayBlockingQueue.java:418)
> at
> org.apache.zookeeper.server.quorum.QuorumCnxManager.pollSendQueue(QuorumCnxManager.java:1294)
> at
> org.apache.zookeeper.server.quorum.QuorumCnxManager.access$700(QuorumCnxManager.java:82)
> at
> org.apache.zookeeper.server.quorum.QuorumCnxManager$SendWorker.run(QuorumCnxManager.java:1131)
> 2020-03-11 20:23:35,744 [myid:2] - WARN
> [SendWorker:3:QuorumCnxManager$SendWorker@1153] - Send worker leaving thread
> id 3 my id = 2
> 2020-03-11 20:23:35,745 [myid:2] - WARN
> [RecvWorker:3:QuorumCnxManager$RecvWorker@1230] - Interrupting
> SendWorker{code}
> The only way I can seem to get them to rejoin the quorum is to restart the
> leader.
> However, if I remove server 4 and 5 from the configuration of server 1 or 2
> (so only servers 1, 2, and 3 remain in the configuration file), then they can
> rejoin the quorum fine. Is this expected and am I doing something wrong? Any
> help or explanation would be greatly appreciated. Thank you.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
[jira] [Commented] (ZOOKEEPER-3756) Members failing to rejoin quorum
[
https://issues.apache.org/jira/browse/ZOOKEEPER-3756?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17058986#comment-17058986
]
Dai Shi commented on ZOOKEEPER-3756:
Ah I forgot one very important file:
[^zoo-service.yaml]
And actually looking through this file I found out why the original issue was
happening :facelpalm:
I had a copy/paste error and forgot to update the selectors for zoo-internal-3
and zoo-internal-4 (the peer and leader election ports) when adding the 2 new
members, so they were both still pointing to zoo-2. After fixing this error
things now behave correctly as they do with a 3 member cluster. Sorry so much
to waste your time!
However, there is still the outstanding issue where restarting the leader pod
causes the cluster to be down for around 30 seconds while it restarts. Is this
expected? I was going to spin up a zookeeper cluster outside of kubernetes
today just to confirm if it still experiences the same behavior. This is a big
issue for running in kubernetes because pods and nodes need to be restarted
relatively frequently due to kubernetes upgrades, and our services cannot
tolerate a 30 second downtime to zookeeper. Since this is not the issue
mentioned at the beginning of this issue, I'm happy to close this and open a
new issue if needed.
> Members failing to rejoin quorum
>
>
> Key: ZOOKEEPER-3756
> URL: https://issues.apache.org/jira/browse/ZOOKEEPER-3756
> Project: ZooKeeper
> Issue Type: Improvement
> Components: leaderElection
>Affects Versions: 3.5.6, 3.5.7
>Reporter: Dai Shi
>Assignee: Mate Szalay-Beko
>Priority: Major
> Attachments: Dockerfile, configmap.yaml, docker-entrypoint.sh,
> jmx.yaml, zoo-service.yaml, zookeeper.yaml
>
>
> Not sure if this is the place to ask, please close if it's not.
> I am seeing some behavior that I can't explain since upgrading to 3.5:
> In a 5 member quorum, when server 3 is the leader and each server has this in
> their configuration:
> {code:java}
> server.1=100.71.255.254:2888:3888:participant;2181
> server.2=100.71.255.253:2888:3888:participant;2181
> server.3=100.71.255.252:2888:3888:participant;2181
> server.4=100.71.255.251:2888:3888:participant;2181
> server.5=100.71.255.250:2888:3888:participant;2181{code}
> If servers 1 or 2 are restarted, they fail to rejoin the quorum with this in
> the logs:
> {code:java}
> 2020-03-11 20:23:35,720 [myid:2] - INFO
> [QuorumPeer[myid=2](plain=0.0.0.0:2181)(secure=disabled):QuorumPeer@1175] -
> LOOKING
> 2020-03-11 20:23:35,721 [myid:2] - INFO
> [QuorumPeer[myid=2](plain=0.0.0.0:2181)(secure=disabled):FastLeaderElection@885]
> - New election. My id = 2, proposed zxid=0x1b8005f4bba
> 2020-03-11 20:23:35,733 [myid:2] - INFO
> [WorkerSender[myid=2]:QuorumCnxManager@438] - Have smaller server identifier,
> so dropping the connection: (3, 2)
> 2020-03-11 20:23:35,734 [myid:2] - INFO
> [0.0.0.0/0.0.0.0:3888:QuorumCnxManager$Listener@924] - Received connection
> request 100.126.116.201:36140
> 2020-03-11 20:23:35,735 [myid:2] - INFO
> [WorkerSender[myid=2]:QuorumCnxManager@438] - Have smaller server identifier,
> so dropping the connection: (4, 2)
> 2020-03-11 20:23:35,740 [myid:2] - INFO
> [WorkerSender[myid=2]:QuorumCnxManager@438] - Have smaller server identifier,
> so dropping the connection: (5, 2)
> 2020-03-11 20:23:35,740 [myid:2] - INFO
> [0.0.0.0/0.0.0.0:3888:QuorumCnxManager$Listener@924] - Received connection
> request 100.126.116.201:36142
> 2020-03-11 20:23:35,740 [myid:2] - INFO
> [WorkerReceiver[myid=2]:FastLeaderElection@679] - Notification: 2 (message
> format version), 2 (n.leader), 0x1b8005f4bba (n.zxid), 0x1 (n.round), LOOKING
> (n.state), 2 (n.sid), 0x1b8 (n.peerEPoch), LOOKING (my state)0 (n.config
> version)
> 2020-03-11 20:23:35,742 [myid:2] - WARN
> [SendWorker:3:QuorumCnxManager$SendWorker@1143] - Interrupted while waiting
> for message on queue
> java.lang.InterruptedException
> at
> java.util.concurrent.locks.AbstractQueuedSynchronizer$ConditionObject.reportInterruptAfterWait(AbstractQueuedSynchronizer.java:2014)
> at
> java.util.concurrent.locks.AbstractQueuedSynchronizer$ConditionObject.awaitNanos(AbstractQueuedSynchronizer.java:2088)
> at
> java.util.concurrent.ArrayBlockingQueue.poll(ArrayBlockingQueue.java:418)
> at
> org.apache.zookeeper.server.quorum.QuorumCnxManager.pollSendQueue(QuorumCnxManager.java:1294)
> at
> org.apache.zookeeper.server.quorum.QuorumCnxManager.access$700(QuorumCnxManager.java:82)
> at
> org.apache.zookeeper.server.quorum.QuorumCnxManager$SendWorker.run(QuorumCnxManager.java:1131)
> 2020-03-11 20:23:35,744 [myid:2] - WARN
> [SendWorker:3:QuorumCnxManager$SendWorker@1153] - Send worker leaving thread
> id 3 my id = 2
> 2020-03-11
[jira] [Commented] (ZOOKEEPER-3756) Members failing to rejoin quorum
[
https://issues.apache.org/jira/browse/ZOOKEEPER-3756?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17058962#comment-17058962
]
Dai Shi commented on ZOOKEEPER-3756:
Here are the kubernetes files:
[^zookeeper.yaml]
[^configmap.yaml]
The zookeeper yaml is slightly redacted to remove some more sensitive
information, but it should only be cosmetic.
Here are the docker image build files:
[^Dockerfile]
[^docker-entrypoint.sh]
[^jmx.yaml]
Note that while docker-entrypoint.sh tries to build the config file if it
doesn't exist, I have since updated our kubernetes config to supply a configmap
with the config file, so the entrypoint doesn't need to generate the config
file anymore.
My gut feeling says this should be able to be reproduced in minikube on a
single machine, but I have not tried. Please let me know if you have any
questions regarding these files or our setup.
> Members failing to rejoin quorum
>
>
> Key: ZOOKEEPER-3756
> URL: https://issues.apache.org/jira/browse/ZOOKEEPER-3756
> Project: ZooKeeper
> Issue Type: Improvement
> Components: leaderElection
>Affects Versions: 3.5.6, 3.5.7
>Reporter: Dai Shi
>Assignee: Mate Szalay-Beko
>Priority: Major
> Attachments: Dockerfile, configmap.yaml, docker-entrypoint.sh,
> jmx.yaml, zookeeper.yaml
>
>
> Not sure if this is the place to ask, please close if it's not.
> I am seeing some behavior that I can't explain since upgrading to 3.5:
> In a 5 member quorum, when server 3 is the leader and each server has this in
> their configuration:
> {code:java}
> server.1=100.71.255.254:2888:3888:participant;2181
> server.2=100.71.255.253:2888:3888:participant;2181
> server.3=100.71.255.252:2888:3888:participant;2181
> server.4=100.71.255.251:2888:3888:participant;2181
> server.5=100.71.255.250:2888:3888:participant;2181{code}
> If servers 1 or 2 are restarted, they fail to rejoin the quorum with this in
> the logs:
> {code:java}
> 2020-03-11 20:23:35,720 [myid:2] - INFO
> [QuorumPeer[myid=2](plain=0.0.0.0:2181)(secure=disabled):QuorumPeer@1175] -
> LOOKING
> 2020-03-11 20:23:35,721 [myid:2] - INFO
> [QuorumPeer[myid=2](plain=0.0.0.0:2181)(secure=disabled):FastLeaderElection@885]
> - New election. My id = 2, proposed zxid=0x1b8005f4bba
> 2020-03-11 20:23:35,733 [myid:2] - INFO
> [WorkerSender[myid=2]:QuorumCnxManager@438] - Have smaller server identifier,
> so dropping the connection: (3, 2)
> 2020-03-11 20:23:35,734 [myid:2] - INFO
> [0.0.0.0/0.0.0.0:3888:QuorumCnxManager$Listener@924] - Received connection
> request 100.126.116.201:36140
> 2020-03-11 20:23:35,735 [myid:2] - INFO
> [WorkerSender[myid=2]:QuorumCnxManager@438] - Have smaller server identifier,
> so dropping the connection: (4, 2)
> 2020-03-11 20:23:35,740 [myid:2] - INFO
> [WorkerSender[myid=2]:QuorumCnxManager@438] - Have smaller server identifier,
> so dropping the connection: (5, 2)
> 2020-03-11 20:23:35,740 [myid:2] - INFO
> [0.0.0.0/0.0.0.0:3888:QuorumCnxManager$Listener@924] - Received connection
> request 100.126.116.201:36142
> 2020-03-11 20:23:35,740 [myid:2] - INFO
> [WorkerReceiver[myid=2]:FastLeaderElection@679] - Notification: 2 (message
> format version), 2 (n.leader), 0x1b8005f4bba (n.zxid), 0x1 (n.round), LOOKING
> (n.state), 2 (n.sid), 0x1b8 (n.peerEPoch), LOOKING (my state)0 (n.config
> version)
> 2020-03-11 20:23:35,742 [myid:2] - WARN
> [SendWorker:3:QuorumCnxManager$SendWorker@1143] - Interrupted while waiting
> for message on queue
> java.lang.InterruptedException
> at
> java.util.concurrent.locks.AbstractQueuedSynchronizer$ConditionObject.reportInterruptAfterWait(AbstractQueuedSynchronizer.java:2014)
> at
> java.util.concurrent.locks.AbstractQueuedSynchronizer$ConditionObject.awaitNanos(AbstractQueuedSynchronizer.java:2088)
> at
> java.util.concurrent.ArrayBlockingQueue.poll(ArrayBlockingQueue.java:418)
> at
> org.apache.zookeeper.server.quorum.QuorumCnxManager.pollSendQueue(QuorumCnxManager.java:1294)
> at
> org.apache.zookeeper.server.quorum.QuorumCnxManager.access$700(QuorumCnxManager.java:82)
> at
> org.apache.zookeeper.server.quorum.QuorumCnxManager$SendWorker.run(QuorumCnxManager.java:1131)
> 2020-03-11 20:23:35,744 [myid:2] - WARN
> [SendWorker:3:QuorumCnxManager$SendWorker@1153] - Send worker leaving thread
> id 3 my id = 2
> 2020-03-11 20:23:35,745 [myid:2] - WARN
> [RecvWorker:3:QuorumCnxManager$RecvWorker@1230] - Interrupting
> SendWorker{code}
> The only way I can seem to get them to rejoin the quorum is to restart the
> leader.
> However, if I remove server 4 and 5 from the configuration of server 1 or 2
> (so only servers 1, 2, and 3 remain in the configuration file), then they can
> rejoin the quorum fine. Is this expected and am I doing something
[jira] [Commented] (ZOOKEEPER-3756) Members failing to rejoin quorum
[
https://issues.apache.org/jira/browse/ZOOKEEPER-3756?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17058800#comment-17058800
]
Mate Szalay-Beko commented on ZOOKEEPER-3756:
-
It is strange indeed...
> We have quorumListenOnAllIPs=true in the config.
I think it is the correct config for kubernetes using 3.5+, I would expect this
to work. But it would be nice to see your full config.
The 3.4 config (where you used 0.0.0.0) is expected to not work in 3.5, see
ZOOKEEPER-2164 and ZOOKEEPER-3725 for details. We already have a fix for this,
but it will be shipped only in the next releases in 3.5.8 and 3.6.1.
If you use {{quorumListenOnAllIPs=true}} and you don't use {{0.0.0.0}} in the
configs, then your problem is something else...
Unfortunately I have not much experience with running ZooKeeper in Kubernetes.
But I would be happy to reproduce your situation and take a deeper look inside
the ZooKeeper leader election. If it is possible for you to share your pod
descriptors / Dockerfiles then I would try them out with both 3 and 5 ZK
servers. Also I would try the fix we committed for ZOOKEEPER-2164, maybe that
helps in this case too. (although this seem to me a different case)
Do you think it is something that can be reproduced on a single host machine
with minikube, or it is something you experienced only with a real kubernetes
cluster?
Alternatively you can share your experience on u...@zookeeper.apache.org - I
know a few guys on that mailing list who are operating ZooKeeper with
Kubernetes, maybe they can share their zookeeper config / pod / docker files.
> Members failing to rejoin quorum
>
>
> Key: ZOOKEEPER-3756
> URL: https://issues.apache.org/jira/browse/ZOOKEEPER-3756
> Project: ZooKeeper
> Issue Type: Improvement
> Components: leaderElection
>Affects Versions: 3.5.6, 3.5.7
>Reporter: Dai Shi
>Assignee: Mate Szalay-Beko
>Priority: Major
>
> Not sure if this is the place to ask, please close if it's not.
> I am seeing some behavior that I can't explain since upgrading to 3.5:
> In a 5 member quorum, when server 3 is the leader and each server has this in
> their configuration:
> {code:java}
> server.1=100.71.255.254:2888:3888:participant;2181
> server.2=100.71.255.253:2888:3888:participant;2181
> server.3=100.71.255.252:2888:3888:participant;2181
> server.4=100.71.255.251:2888:3888:participant;2181
> server.5=100.71.255.250:2888:3888:participant;2181{code}
> If servers 1 or 2 are restarted, they fail to rejoin the quorum with this in
> the logs:
> {code:java}
> 2020-03-11 20:23:35,720 [myid:2] - INFO
> [QuorumPeer[myid=2](plain=0.0.0.0:2181)(secure=disabled):QuorumPeer@1175] -
> LOOKING
> 2020-03-11 20:23:35,721 [myid:2] - INFO
> [QuorumPeer[myid=2](plain=0.0.0.0:2181)(secure=disabled):FastLeaderElection@885]
> - New election. My id = 2, proposed zxid=0x1b8005f4bba
> 2020-03-11 20:23:35,733 [myid:2] - INFO
> [WorkerSender[myid=2]:QuorumCnxManager@438] - Have smaller server identifier,
> so dropping the connection: (3, 2)
> 2020-03-11 20:23:35,734 [myid:2] - INFO
> [0.0.0.0/0.0.0.0:3888:QuorumCnxManager$Listener@924] - Received connection
> request 100.126.116.201:36140
> 2020-03-11 20:23:35,735 [myid:2] - INFO
> [WorkerSender[myid=2]:QuorumCnxManager@438] - Have smaller server identifier,
> so dropping the connection: (4, 2)
> 2020-03-11 20:23:35,740 [myid:2] - INFO
> [WorkerSender[myid=2]:QuorumCnxManager@438] - Have smaller server identifier,
> so dropping the connection: (5, 2)
> 2020-03-11 20:23:35,740 [myid:2] - INFO
> [0.0.0.0/0.0.0.0:3888:QuorumCnxManager$Listener@924] - Received connection
> request 100.126.116.201:36142
> 2020-03-11 20:23:35,740 [myid:2] - INFO
> [WorkerReceiver[myid=2]:FastLeaderElection@679] - Notification: 2 (message
> format version), 2 (n.leader), 0x1b8005f4bba (n.zxid), 0x1 (n.round), LOOKING
> (n.state), 2 (n.sid), 0x1b8 (n.peerEPoch), LOOKING (my state)0 (n.config
> version)
> 2020-03-11 20:23:35,742 [myid:2] - WARN
> [SendWorker:3:QuorumCnxManager$SendWorker@1143] - Interrupted while waiting
> for message on queue
> java.lang.InterruptedException
> at
> java.util.concurrent.locks.AbstractQueuedSynchronizer$ConditionObject.reportInterruptAfterWait(AbstractQueuedSynchronizer.java:2014)
> at
> java.util.concurrent.locks.AbstractQueuedSynchronizer$ConditionObject.awaitNanos(AbstractQueuedSynchronizer.java:2088)
> at
> java.util.concurrent.ArrayBlockingQueue.poll(ArrayBlockingQueue.java:418)
> at
> org.apache.zookeeper.server.quorum.QuorumCnxManager.pollSendQueue(QuorumCnxManager.java:1294)
> at
> org.apache.zookeeper.server.quorum.QuorumCnxManager.access$700(QuorumCnxManager.java:82)
> at
>
[jira] [Commented] (ZOOKEEPER-3756) Members failing to rejoin quorum
[
https://issues.apache.org/jira/browse/ZOOKEEPER-3756?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17058718#comment-17058718
]
Dai Shi commented on ZOOKEEPER-3756:
Hi Mate,
Thanks for taking a look at this! I suspected it may have something to do with
the IPs, so hoping you may be able to explain more clearly what might be
happening :)
We are running zookeeper inside kubernetes as a stateful set. The IP
100.126.116.201 is the actual IP of the pod that the leader, member 3, is
running in. Unfortunately as far as I know you cannot set static IPs for pods
in kubernetes, so that is why the IPs in the configs do not match the IP that
zookeeper binds to since we cannot know in advance what the pod IP will be. We
have quorumListenOnAllIPs=true in the config. The IPs in the config point to
kubernetes services, which remain static, each of which select the single
corresponding pod as the backend.
Just for reference, prior to upgrading to 3.5 we were using 0.0.0.0 in our
configs, but only in a 3 member cluster. As far as I can tell, there are no
issues with member rejoining the quorum with either the 3.4 config or 3.5
config in a 3 member cluster. However, there is one major issue in both that if
the leader pod is restarted, the cluster becomes hard down until the pod
restarts. It seems that the leader election does not take place until the pod
comes back and the zookeeper process starts, so even though the logs say leader
election took 1ms, the cluster was down for maybe 30 to 60 seconds.
This is what led to me trying a 5 member cluster to see if that would trigger
the leader election quicker, but it led to me discovering this behavior
instead. I still don't understand why if I remove servers 4 and 5 from the
config of servers 1 and 2 they can then successfully rejoin.
I've since shifted the cluster back down to 3 members, but if you think it
would be helpful to get a log from all 5 members I can repeat the experiment
again.
Thanks again for your help!
> Members failing to rejoin quorum
>
>
> Key: ZOOKEEPER-3756
> URL: https://issues.apache.org/jira/browse/ZOOKEEPER-3756
> Project: ZooKeeper
> Issue Type: Improvement
> Components: leaderElection
>Affects Versions: 3.5.6, 3.5.7
>Reporter: Dai Shi
>Assignee: Mate Szalay-Beko
>Priority: Major
>
> Not sure if this is the place to ask, please close if it's not.
> I am seeing some behavior that I can't explain since upgrading to 3.5:
> In a 5 member quorum, when server 3 is the leader and each server has this in
> their configuration:
> {code:java}
> server.1=100.71.255.254:2888:3888:participant;2181
> server.2=100.71.255.253:2888:3888:participant;2181
> server.3=100.71.255.252:2888:3888:participant;2181
> server.4=100.71.255.251:2888:3888:participant;2181
> server.5=100.71.255.250:2888:3888:participant;2181{code}
> If servers 1 or 2 are restarted, they fail to rejoin the quorum with this in
> the logs:
> {code:java}
> 2020-03-11 20:23:35,720 [myid:2] - INFO
> [QuorumPeer[myid=2](plain=0.0.0.0:2181)(secure=disabled):QuorumPeer@1175] -
> LOOKING
> 2020-03-11 20:23:35,721 [myid:2] - INFO
> [QuorumPeer[myid=2](plain=0.0.0.0:2181)(secure=disabled):FastLeaderElection@885]
> - New election. My id = 2, proposed zxid=0x1b8005f4bba
> 2020-03-11 20:23:35,733 [myid:2] - INFO
> [WorkerSender[myid=2]:QuorumCnxManager@438] - Have smaller server identifier,
> so dropping the connection: (3, 2)
> 2020-03-11 20:23:35,734 [myid:2] - INFO
> [0.0.0.0/0.0.0.0:3888:QuorumCnxManager$Listener@924] - Received connection
> request 100.126.116.201:36140
> 2020-03-11 20:23:35,735 [myid:2] - INFO
> [WorkerSender[myid=2]:QuorumCnxManager@438] - Have smaller server identifier,
> so dropping the connection: (4, 2)
> 2020-03-11 20:23:35,740 [myid:2] - INFO
> [WorkerSender[myid=2]:QuorumCnxManager@438] - Have smaller server identifier,
> so dropping the connection: (5, 2)
> 2020-03-11 20:23:35,740 [myid:2] - INFO
> [0.0.0.0/0.0.0.0:3888:QuorumCnxManager$Listener@924] - Received connection
> request 100.126.116.201:36142
> 2020-03-11 20:23:35,740 [myid:2] - INFO
> [WorkerReceiver[myid=2]:FastLeaderElection@679] - Notification: 2 (message
> format version), 2 (n.leader), 0x1b8005f4bba (n.zxid), 0x1 (n.round), LOOKING
> (n.state), 2 (n.sid), 0x1b8 (n.peerEPoch), LOOKING (my state)0 (n.config
> version)
> 2020-03-11 20:23:35,742 [myid:2] - WARN
> [SendWorker:3:QuorumCnxManager$SendWorker@1143] - Interrupted while waiting
> for message on queue
> java.lang.InterruptedException
> at
> java.util.concurrent.locks.AbstractQueuedSynchronizer$ConditionObject.reportInterruptAfterWait(AbstractQueuedSynchronizer.java:2014)
> at
>
[jira] [Commented] (ZOOKEEPER-3756) Members failing to rejoin quorum
[
https://issues.apache.org/jira/browse/ZOOKEEPER-3756?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17058642#comment-17058642
]
Mate Szalay-Beko commented on ZOOKEEPER-3756:
-
Hello!
I was working on these parts recently and happy to take a look on your case.
The log file you sent is more-or-less OK. In ZooKeeper, the servers are
communicating with each other using the 3888 port (in your config) for election
protocol. When a server starts, it tries to connect to all other server's
election port and asks for the IDs from each server. But only those channels
remain, which was initiated by the servers with higher ID. This is why you see
the following message, which is completely normal:
{code:java}
2020-03-11 20:23:35,733 [myid:2] - INFO
[WorkerSender[myid=2]:QuorumCnxManager@438] - Have smaller server identifier,
so dropping the connection: (3, 2)
{code}
And also the SendWorker thread gets interrupted towards server 3 because of the
same reason. That is also OK.
The only strange thing I noticed in your logs are in these lines:
{code:java}
2020-03-11 20:23:35,734 [myid:2] - INFO
[0.0.0.0/0.0.0.0:3888:QuorumCnxManager$Listener@924] - Received connection
request 100.126.116.201:36140
2020-03-11 20:23:35,740 [myid:2] - INFO
[0.0.0.0/0.0.0.0:3888:QuorumCnxManager$Listener@924] - Received connection
request 100.126.116.201:36142
{code}
It looks like you get two connection requests from this IP: 100.126.116.201.
This IP is not in your config.
Are you sure you are using the same config you sent in all ZK nodes? Can you
check where this 100.126.116.201 comes from?
There is a known bug which could explain your situation, but that would happen
only if you use 0.0.0.0 in your configs.
Are you using some dockerized environment maybe?
Can you share the ZooKeeper configs and server logs from all the 5 nodes?
> Members failing to rejoin quorum
>
>
> Key: ZOOKEEPER-3756
> URL: https://issues.apache.org/jira/browse/ZOOKEEPER-3756
> Project: ZooKeeper
> Issue Type: Improvement
> Components: leaderElection
>Affects Versions: 3.5.6, 3.5.7
>Reporter: Dai Shi
>Assignee: Mate Szalay-Beko
>Priority: Major
>
> Not sure if this is the place to ask, please close if it's not.
> I am seeing some behavior that I can't explain since upgrading to 3.5:
> In a 5 member quorum, when server 3 is the leader and each server has this in
> their configuration:
> {code:java}
> server.1=100.71.255.254:2888:3888:participant;2181
> server.2=100.71.255.253:2888:3888:participant;2181
> server.3=100.71.255.252:2888:3888:participant;2181
> server.4=100.71.255.251:2888:3888:participant;2181
> server.5=100.71.255.250:2888:3888:participant;2181{code}
> If servers 1 or 2 are restarted, they fail to rejoin the quorum with this in
> the logs:
> {code:java}
> 2020-03-11 20:23:35,720 [myid:2] - INFO
> [QuorumPeer[myid=2](plain=0.0.0.0:2181)(secure=disabled):QuorumPeer@1175] -
> LOOKING
> 2020-03-11 20:23:35,721 [myid:2] - INFO
> [QuorumPeer[myid=2](plain=0.0.0.0:2181)(secure=disabled):FastLeaderElection@885]
> - New election. My id = 2, proposed zxid=0x1b8005f4bba
> 2020-03-11 20:23:35,733 [myid:2] - INFO
> [WorkerSender[myid=2]:QuorumCnxManager@438] - Have smaller server identifier,
> so dropping the connection: (3, 2)
> 2020-03-11 20:23:35,734 [myid:2] - INFO
> [0.0.0.0/0.0.0.0:3888:QuorumCnxManager$Listener@924] - Received connection
> request 100.126.116.201:36140
> 2020-03-11 20:23:35,735 [myid:2] - INFO
> [WorkerSender[myid=2]:QuorumCnxManager@438] - Have smaller server identifier,
> so dropping the connection: (4, 2)
> 2020-03-11 20:23:35,740 [myid:2] - INFO
> [WorkerSender[myid=2]:QuorumCnxManager@438] - Have smaller server identifier,
> so dropping the connection: (5, 2)
> 2020-03-11 20:23:35,740 [myid:2] - INFO
> [0.0.0.0/0.0.0.0:3888:QuorumCnxManager$Listener@924] - Received connection
> request 100.126.116.201:36142
> 2020-03-11 20:23:35,740 [myid:2] - INFO
> [WorkerReceiver[myid=2]:FastLeaderElection@679] - Notification: 2 (message
> format version), 2 (n.leader), 0x1b8005f4bba (n.zxid), 0x1 (n.round), LOOKING
> (n.state), 2 (n.sid), 0x1b8 (n.peerEPoch), LOOKING (my state)0 (n.config
> version)
> 2020-03-11 20:23:35,742 [myid:2] - WARN
> [SendWorker:3:QuorumCnxManager$SendWorker@1143] - Interrupted while waiting
> for message on queue
> java.lang.InterruptedException
> at
> java.util.concurrent.locks.AbstractQueuedSynchronizer$ConditionObject.reportInterruptAfterWait(AbstractQueuedSynchronizer.java:2014)
> at
> java.util.concurrent.locks.AbstractQueuedSynchronizer$ConditionObject.awaitNanos(AbstractQueuedSynchronizer.java:2088)
> at
> java.util.concurrent.ArrayBlockingQueue.poll(ArrayBlockingQueue.java:418)
> at
>
