[
https://issues.apache.org/jira/browse/KAFKA-4277?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15950689#comment-15950689
]
Johan Ström commented on KAFKA-4277:
------------------------------------
Hi,
we just had the same problem.
Our setup:
4x kafka_2.11-0.10.1.1
3x zookeeper-3.4.8
For now, only 1 topic + consumer_offsets, pretty low traffic and fortunately no
critical data.
For some external reason (not yet known), the three zookeepers lost connection
with each other, and the brokers all lost connection to Zookeeper. This
happened twice.
All nodes reported they where in state Running via kafka.metrics.reporters
(kafka.metrics.KafkaGraphiteMetricsReporter in our case). Some did however
report underreplicated partitions, that was how we noticed it in the first
place (unfortunately too late by watching metrics, no active monitoring of the
actual numbers).
The two Zk followers logged Read timeout at exactly the same time:
zk1: 2017-03-28 23:05:55,923
zk2: 2017-03-28 23:05:55,923
zk3: time unknown, log was rolled over
Within a second they are back online, now with 2 as leader.
Almost an hour later, a new drop:
{noformat}
2017-03-28 23:59:08,901 [myid:1] - WARN
[QuorumPeer[myid=1]/0.0.0.0:2181:Follower@89] - Exception when following the
leader
java.net.SocketTimeoutException: Read timed out
{noformat}
zk2 (now leader) reported a similar:
{noformat}
2017-03-28 23:59:08,901 [myid:2] - ERROR
[LearnerHandler-/10.88.176.158:46722:LearnerHandler@631] - Unexpected exception
causing shutdown while sock still open
java.net.SocketTimeoutException: Read timed out
{noformat}
Unfortunately no log from zk3 due to rollover. After this, zk3 took over as
leader.
Let's first look at where zk2 took over as leader. Immediately after, it
started logging these:
{noformat}
2017-03-28 23:05:56,671 [myid:2] - INFO [ProcessThread(sid:2
cport:-1)::PrepRequestProcessor@651] - Got user-level KeeperException when
processing sessionid:0x15b16bd4cd00000 type:create cxid:0x3 zxid:0x900000002
txntype:-1 reqpath:n/a Error Path:/kafka/brokers Error:KeeperErrorCode =
NodeExists for /kafka/brokers
2017-03-28 23:05:56,671 [myid:2] - INFO [ProcessThread(sid:2
cport:-1)::PrepRequestProcessor@651] - Got user-level KeeperException when
processing sessionid:0x15b16bd4cd00000 type:create cxid:0x4 zxid:0x900000003
txntype:-1 reqpath:n/a Error Path:/kafka/brokers/ids Error:KeeperErrorCode =
NodeExists for /kafka/brokers/ids
2017-03-28 23:05:56,672 [myid:2] - INFO [ProcessThread(sid:2
cport:-1)::PrepRequestProcessor@651] - Got user-level KeeperException when
processing sessionid:0x15b16bd4cd00000 type:create cxid:0x5 zxid:0x900000004
txntype:-1 reqpath:n/a Error Path:/kafka/brokers/ids/2 Error:KeeperErrorCode =
NodeExists for /kafka/brokers/ids/2
{noformat}
Unfortunately, this never recovers and /brokers/ids/ are eventually empty,
since Kafka fails to re-register.
In addition to the above, it also starts spamming these:
{noformat}
2017-03-28 23:06:19,611 [myid:2] - INFO [ProcessThread(sid:2
cport:-1)::PrepRequestProcessor@651] - Got user-level KeeperException when
processing sessionid:0x15b16bd4cd00001 type:setData cxid:0xb zxid:0x900000077
txntype:-1 re
qpath:n/a Error
Path:/kafka/brokers/topics/iptv-engagement-event/partitions/3/state
Error:KeeperErrorCode = BadVersion for
/kafka/brokers/topics/iptv-engagement-event/partitions/3/state
{noformat}
These will continue for all partitions (or at least alot of them) on the leader
(first 2 then 3) until we restart the brokers.
---
When it comes to the brokers. broker 1, at first error:
{noformat}
[2017-03-28 23:05:55,560] WARN Client session timed out, have not heard from
server in 4002ms for sessionid 0x359b1ad98b00004
(org.apache.zookeeper.ClientCnxn)
[2017-03-28 23:05:55,560] INFO Client session timed out, have not heard from
server in 4002ms for sessionid 0x359b1ad98b00004, closing socket connection and
attempting reconnect (org.apache.zookeeper.ClientCnxn)
[2017-03-28 23:05:55,661] INFO zookeeper state changed (Disconnected)
(org.I0Itec.zkclient.ZkClient)
[2017-03-28 23:05:55,711] INFO Opening socket connection to server
zookeeper02/10.88.176.168:2181. Will not attempt to authenticate using SASL
(unknown error) (org.apache.zookeeper.ClientCnxn)
[2017-03-28 23:05:56,711] INFO Socket connection established to
zookeeper02/10.88.176.168:2181, initiating session
(org.apache.zookeeper.ClientCnxn)
[2017-03-28 23:05:56,714] INFO Session establishment complete on server
zookeeper02/10.88.176.168:2181, sessionid = 0x359b1ad98b00004, negotiated
timeout = 6000 (org.apache.zookeeper.ClientCnxn)
[2017-03-28 23:05:56,714] INFO zookeeper state changed (SyncConnected)
(org.I0Itec.zkclient.ZkClient)
{noformat}
After this, it dropped some replication fetchers but no more errors. Until the
next outage:
Almost one hour later, a new timeout (still broker 1):
{noformat}
[2017-03-28 23:59:01,596] WARN Client session timed out, have not heard from
server in 4002ms for sessionid 0x359b1ad98b00004
(org.apache.zookeeper.ClientCnxn)
[2017-03-28 23:59:01,597] INFO Client session timed out, have not heard from
server in 4002ms for sessionid 0x359b1ad98b00004, closing socket connection and
attempting reconnect (org.apache.zookeeper.ClientCnxn)
[2017-03-28 23:59:01,697] INFO zookeeper state changed (Disconnected)
(org.I0Itec.zkclient.ZkClient)
[2017-03-28 23:59:01,736] INFO Opening socket connection to server
zookeeper01/10.88.176.167:2181. Will not attempt to authenticate using SASL
(unknown error) (org.apache.zookeeper.ClientCnxn)
[2017-03-28 23:59:03,739] WARN Client session timed out, have not heard from
server in 2042ms for sessionid 0x359b1ad98b00004
(org.apache.zookeeper.ClientCnxn)
[2017-03-28 23:59:03,739] INFO Client session timed out, have not heard from
server in 2042ms for sessionid 0x359b1ad98b00004, closing socket connection and
attempting reconnect (org.apache.zookeeper.ClientCnxn)
[2017-03-28 23:59:04,375] INFO Opening socket connection to server
zookeeper03/10.88.176.158:2181. Will not attempt to authenticate using SASL
(unknown error) (org.apache.zookeeper.ClientCnxn)
[2017-03-28 23:59:06,376] WARN Client session timed out, have not heard from
server in 2537ms for sessionid 0x359b1ad98b00004
(org.apache.zookeeper.ClientCnxn)
[2017-03-28 23:59:06,376] INFO Client session timed out, have not heard from
server in 2537ms for sessionid 0x359b1ad98b00004, closing socket connection and
attempting reconnect (org.apache.zookeeper.ClientCnxn)
[2017-03-28 23:59:07,849] INFO Opening socket connection to server
zookeeper02/10.88.176.168:2181. Will not attempt to authenticate using SASL
(unknown error) (org.apache.zookeeper.ClientCnxn)
[2017-03-28 23:59:07,849] INFO Socket connection established to
zookeeper02/10.88.176.168:2181, initiating session
(org.apache.zookeeper.ClientCnxn)
[2017-03-28 23:59:07,850] WARN Unable to reconnect to ZooKeeper service,
session 0x359b1ad98b00004 has expired (org.apache.zookeeper.ClientCnxn)
[2017-03-28 23:59:07,850] INFO zookeeper state changed (Expired)
(org.I0Itec.zkclient.ZkClient)
[2017-03-28 23:59:07,850] INFO Initiating client connection,
connectString=zookeeper01:2181,zookeeper02:2181,zookeeper03:2181/kafka
sessionTimeout=6000 watcher=org.I0Itec.zkclient.ZkClient@4fb61f4a
(org.apache.zookeeper.ZooKeeper)
[2017-03-28 23:59:07,850] INFO Unable to reconnect to ZooKeeper service,
session 0x359b1ad98b00004 has expired, closing socket connection
(org.apache.zookeeper.ClientCnxn)
[2017-03-28 23:59:07,852] INFO Opening socket connection to server
zookeeper02/10.88.176.168:2181. Will not attempt to authenticate using SASL
(unknown error) (org.apache.zookeeper.ClientCnxn)
[2017-03-28 23:59:07,852] INFO EventThread shut down for session:
0x359b1ad98b00004 (org.apache.zookeeper.ClientCnxn)
[2017-03-28 23:59:07,852] INFO Socket connection established to
zookeeper02/10.88.176.168:2181, initiating session
(org.apache.zookeeper.ClientCnxn)
[2017-03-28 23:59:09,854] WARN Client session timed out, have not heard from
server in 2002ms for sessionid 0x0 (org.apache.zookeeper.ClientCnxn)
[2017-03-28 23:59:09,854] INFO Client session timed out, have not heard from
server in 2002ms for sessionid 0x0, closing socket connection and attempting
reconnect (org.apache.zookeeper.ClientCnxn)
[2017-03-28 23:59:09,955] INFO Waiting for keeper state SyncConnected
(org.I0Itec.zkclient.ZkClient)
[2017-03-28 23:59:10,699] INFO Opening socket connection to server
zookeeper01/10.88.176.167:2181. Will not attempt to authenticate using SASL
(unknown error) (org.apache.zookeeper.ClientCnxn)
[2017-03-28 23:59:10,700] INFO Socket connection established to
zookeeper01/10.88.176.167:2181, initiating session
(org.apache.zookeeper.ClientCnxn)
[2017-03-28 23:59:10,702] INFO Session establishment complete on server
zookeeper01/10.88.176.167:2181, sessionid = 0x15b16ee049c0000, negotiated
timeout = 6000 (org.apache.zookeeper.ClientCnxn)
[2017-03-28 23:59:10,702] INFO zookeeper state changed (SyncConnected)
(org.I0Itec.zkclient.ZkClient)
[2017-03-28 23:59:10,705] INFO re-registering broker info in ZK for broker 1
(kafka.server.KafkaHealthcheck$SessionExpireListener)
[2017-03-28 23:59:10,706] INFO Creating /brokers/ids/1 (is it secure? false)
(kafka.utils.ZKCheckedEphemeral)
[2017-03-28 23:59:10,709] INFO Result of znode creation is: NODEEXISTS
(kafka.utils.ZKCheckedEphemeral)
[2017-03-28 23:59:10,709] ERROR Error handling event ZkEvent[New session event
sent to kafka.server.KafkaHealthcheck$SessionExpireListener@70bba1f7]
(org.I0Itec.zkclient.ZkEventThread)
java.lang.RuntimeException: A broker is already registered on the path
/brokers/ids/1. This probably indicates that you either have configured a
brokerid that is already in use, or else you have shutdown this broker and
restarted it faster than the zookeeper timeout so it appears to be
re-registering.
at kafka.utils.ZkUtils.registerBrokerInZk(ZkUtils.scala:393)
at kafka.utils.ZkUtils.registerBrokerInZk(ZkUtils.scala:379)
at kafka.server.KafkaHealthcheck.register(KafkaHealthcheck.scala:70)
at
kafka.server.KafkaHealthcheck$SessionExpireListener.handleNewSession(KafkaHealthcheck.scala:104)
at org.I0Itec.zkclient.ZkClient$6.run(ZkClient.java:735)
at org.I0Itec.zkclient.ZkEventThread.run(ZkEventThread.java:71)
{noformat}
Comparing to the other brokers, broker 2 and 3 failed its /brokers/ids/N
registration at 23:05:55, and broker 4 failed at 23:59:11 same as broker 1.
Let's take a look at broker1's (now expired) ZK session with ID
0x359b1ad98b00004:
{noformat}
[2017-03-28 23:05:56,714] INFO Session establishment complete on server
zookeeper02/10.88.176.168:2181, sessionid = 0x359b1ad98b00004, negotiated
timeout = 6000 (org.apache.zookeeper.ClientCnxn)
...
[2017-03-28 23:59:01,596] WARN Client session timed out, have not heard from
server in 4002ms for sessionid 0x359b1ad98b00004
(org.apache.zookeeper.ClientCnxn)
[2017-03-28 23:59:01,597] INFO Client session timed out, have not heard from
server in 4002ms for sessionid 0x359b1ad98b00004, closing socket connection and
attempting reconnect (org.apache.zookeeper.ClientCnxn)
[2017-03-28 23:59:01,697] INFO zookeeper state changed (Disconnected)
(org.I0Itec.zkclient.ZkClient)
[2017-03-28 23:59:01,736] INFO Opening socket connection to server
zookeeper01/10.88.176.167:2181. Will not attempt to authenticate using SASL
(unknown error) (org.apache.zookeeper.ClientCnxn)
[2017-03-28 23:59:03,739] WARN Client session timed out, have not heard from
server in 2042ms for sessionid 0x359b1ad98b00004
(org.apache.zookeeper.ClientCnxn)
[2017-03-28 23:59:03,739] INFO Client session timed out, have not heard from
server in 2042ms for sessionid 0x359b1ad98b00004, closing socket connection and
attempting reconnect (org.apache.zookeeper.ClientCnxn)
[2017-03-28 23:59:04,375] INFO Opening socket connection to server
zookeeper03/10.88.176.158:2181. Will not attempt to authenticate using SASL
(unknown error) (org.apache.zookeeper.ClientCnxn)
[2017-03-28 23:59:06,376] WARN Client session timed out, have not heard from
server in 2537ms for sessionid 0x359b1ad98b00004
(org.apache.zookeeper.ClientCnxn)
[2017-03-28 23:59:06,376] INFO Client session timed out, have not heard from
server in 2537ms for sessionid 0x359b1ad98b00004, closing socket connection and
attempting reconnect (org.apache.zookeeper.ClientCnxn)
[2017-03-28 23:59:07,849] INFO Opening socket connection to server
zookeeper02/10.88.176.168:2181. Will not attempt to authenticate using SASL
(unknown error) (org.apache.zookeeper.ClientCnxn)
[2017-03-28 23:59:07,849] INFO Socket connection established to
zookeeper02/10.88.176.168:2181, initiating session
(org.apache.zookeeper.ClientCnxn)
[2017-03-28 23:59:07,850] WARN Unable to reconnect to ZooKeeper service,
session 0x359b1ad98b00004 has expired (org.apache.zookeeper.ClientCnxn)
[2017-03-28 23:59:07,850] INFO zookeeper state changed (Expired)
(org.I0Itec.zkclient.ZkClient)
[2017-03-28 23:59:07,850] INFO Initiating client connection,
connectString=zookeeper01:2181,zookeeper02:2181,zookeeper03:2181/kafka
sessionTimeout=6000 watcher=org.I0Itec.zkclient.ZkClient@4fb61f4a
(org.apache.zookeeper.ZooKeeper)
[2017-03-28 23:59:07,850] INFO Unable to reconnect to ZooKeeper service,
session 0x359b1ad98b00004 has expired, closing socket connection
(org.apache.zookeeper.ClientCnxn)
[2017-03-28 23:59:07,852] INFO Opening socket connection to server
zookeeper02/10.88.176.168:2181. Will not attempt to authenticate using SASL
(unknown error) (org.apache.zookeeper.ClientCnxn)
[2017-03-28 23:59:07,852] INFO EventThread shut down for session:
0x359b1ad98b00004 (org.apache.zookeeper.ClientCnxn)
{noformat}
The same session, on the zookeeper side:
{noformat}
2017-03-28 23:05:56,714 [myid:2] - INFO
[NIOServerCxn.Factory:0.0.0.0/0.0.0.0:2181:ZooKeeperServer@893] - Client
attempting to renew session 0x359b1ad98b00004 at /10.88.176.164:60633
2017-03-28 23:05:56,714 [myid:2] - INFO
[NIOServerCxn.Factory:0.0.0.0/0.0.0.0:2181:ZooKeeperServer@645] - Established
session 0x359b1ad98b00004 with negotiated timeout 6000 for client
/10.88.176.164:60633
...
2017-03-28 23:59:04,000 [myid:2] - INFO [SessionTracker:ZooKeeperServer@355] -
Expiring session 0x359b1ad98b00004, timeout of 6000ms exceeded
2017-03-28 23:59:04,001 [myid:2] - INFO [ProcessThread(sid:2
cport:-1)::PrepRequestProcessor@489] - Processed session termination for
sessionid: 0x359b1ad98b00004
2017-03-28 23:59:05,867 [myid:2] - WARN
[NIOServerCxn.Factory:0.0.0.0/0.0.0.0:2181:NIOServerCnxn@357] - caught end of
stream exception
EndOfStreamException: Unable to read additional data from client sessionid
0x359b1ad98b00004, likely client has closed socket
at
org.apache.zookeeper.server.NIOServerCnxn.doIO(NIOServerCnxn.java:230)
at
org.apache.zookeeper.server.NIOServerCnxnFactory.run(NIOServerCnxnFactory.java:203)
at java.lang.Thread.run(Thread.java:745)
2017-03-28 23:59:05,867 [myid:2] - INFO
[NIOServerCxn.Factory:0.0.0.0/0.0.0.0:2181:NIOServerCnxn@1008] - Closed socket
connection for client /10.88.176.164:60633 which had sessionid 0x359b1ad98b00004
...
2017-03-28 23:59:07,850 [myid:2] - INFO
[NIOServerCxn.Factory:0.0.0.0/0.0.0.0:2181:ZooKeeperServer@893] - Client
attempting to renew session 0x359b1ad98b00004 at /10.88.176.164:32922
2017-03-28 23:59:07,850 [myid:2] - INFO
[NIOServerCxn.Factory:0.0.0.0/0.0.0.0:2181:ZooKeeperServer@638] - Invalid
session 0x359b1ad98b00004 for client /10.88.176.164:32922, probably expired
2017-03-28 23:59:07,850 [myid:2] - INFO
[NIOServerCxn.Factory:0.0.0.0/0.0.0.0:2181:NIOServerCnxn@1008] - Closed socket
connection for client /10.88.176.164:32922 which had sessionid 0x359b1ad98b00004
{noformat}
So, basically, the brokers are never able to re-register as described by others
in this ticket.
broker 1 doesn't log any more errors at all, and neither does broker 4.
But, on broker 2 and broker 3, the following is spammed for all (most)
partitions until they are restarted:
{noformat}
[2017-03-28 23:05:56,674] ERROR Error handling event ZkEvent[New session event
sent to kafka.server.KafkaHealthcheck$SessionExpireListener@3767aad0]
(org.I0Itec.zkclient.ZkEventThread)
java.lang.RuntimeException: A broker is already registered on the path
/brokers/ids/2. This probably indicates that you either have configured a
brokerid that is already in use, or else you have shutdown this broker and
restarted it faster than the zookeeper timeout so it appears to be
re-registering.
[2017-03-28 23:06:20,454] INFO Partition [__consumer_offsets,37] on broker 2:
Shrinking ISR for partition [__consumer_offsets,37] from 2,1,3 to 2,3
(kafka.cluster.Partition)
[2017-03-28 23:06:20,466] INFO Partition [__consumer_offsets,37] on broker 2:
Cached zkVersion [0] not equal to that in zookeeper, skip updating ISR
(kafka.cluster.Partition)
..... and one it goes forever
{noformat}
Hopefully this can help pinpoint the cause.. and get some fix done.. Let me
know if more logs are useful, I'll keep them (except for zookeeper 3 which was
overwritten before this was noticed).
> creating ephemeral node already exist
> -------------------------------------
>
> Key: KAFKA-4277
> URL: https://issues.apache.org/jira/browse/KAFKA-4277
> Project: Kafka
> Issue Type: Bug
> Affects Versions: 0.10.0.0
> Reporter: Feixiang Yan
>
> I use zookeeper 3.4.6.
> Zookeeper session time out, zkClient try reconnect failed. Then re-establish
> the session and re-registering broker info in ZK, throws NODEEXISTS Exception.
> I think it is because the ephemeral node which created by old session has
> not removed.
> I read the
> [ZkUtils.scala|https://github.com/apache/kafka/blob/0.8.1/core/src/main/scala/kafka/utils/ZkUtils.scala]
> of 0.8.1, createEphemeralPathExpectConflictHandleZKBug try create node in a
> while loop until create success. This can solve the issue. But in
> [ZkUtils.scala|https://github.com/apache/kafka/blob/0.10.0.1/core/src/main/scala/kafka/utils/ZkUtils.scala]
> 0.10.1 the function removed.
> {noformat}
> [2016-10-07 19:00:32,562] INFO Socket connection established to
> 10.191.155.238/10.191.155.238:21819, initiating session
> (org.apache.zookeeper.ClientCnxn)
> [2016-10-07 19:00:32,563] INFO zookeeper state changed (Expired)
> (org.I0Itec.zkclient.ZkClient)
> [2016-10-07 19:00:32,564] INFO Unable to reconnect to ZooKeeper service,
> session 0x1576b11f9b201bd has expired, closing socket connection
> (org.apache.zookeeper.ClientCnxn)
> [2016-10-07 19:00:32,564] INFO Initiating client connection,
> connectString=10.191.155.237:21819,10.191.155.238:21819,10.191.155.239:21819/cluster2
> sessionTimeout=6000 watcher=org.I0Itec.zkclient.ZkClient@ae71be2
> (org.apache.zookeeper.ZooKeeper)
> [2016-10-07 19:00:32,566] INFO Opening socket connection to server
> 10.191.155.237/10.191.155.237:21819. Will not attempt to authenticate using
> SASL (unknown error) (org.apache.zookeeper.ClientCnxn)
> [2016-10-07 19:00:32,566] INFO Socket connection established to
> 10.191.155.237/10.191.155.237:21819, initiating session
> (org.apache.zookeeper.ClientCnxn)
> [2016-10-07 19:00:32,566] INFO EventThread shut down
> (org.apache.zookeeper.ClientCnxn)
> [2016-10-07 19:00:32,567] INFO Session establishment complete on server
> 10.191.155.237/10.191.155.237:21819, sessionid = 0x1579ecd39c20006,
> negotiated timeout = 6000 (org.apache.zookeeper.ClientCnxn)
> [2016-10-07 19:00:32,567] INFO zookeeper state changed (SyncConnected)
> (org.I0Itec.zkclient.ZkClient)
> [2016-10-07 19:00:32,608] INFO re-registering broker info in ZK for broker 3
> (kafka.server.KafkaHealthcheck$SessionExpireListener)
> [2016-10-07 19:00:32,610] INFO Creating /brokers/ids/3 (is it secure? false)
> (kafka.utils.ZKCheckedEphemeral)
> [2016-10-07 19:00:32,611] INFO Result of znode creation is: NODEEXISTS
> (kafka.utils.ZKCheckedEphemeral)
> [2016-10-07 19:00:32,614] ERROR Error handling event ZkEvent[New session
> event sent to kafka.server.KafkaHealthcheck$SessionExpireListener@324f1bc]
> (org.I0Itec.zkclient.ZkEventThread)
> java.lang.RuntimeException: A broker is already registered on the path
> /brokers/ids/3. This probably indicates that you either have configured a
> brokerid that is already in use, or else you have shutdown this broker and
> restarted it faster than the zookeeper timeout so it appears to be
> re-registering.
> at kafka.utils.ZkUtils.registerBrokerInZk(ZkUtils.scala:305)
> at kafka.utils.ZkUtils.registerBrokerInZk(ZkUtils.scala:291)
> at kafka.server.KafkaHealthcheck.register(KafkaHealthcheck.scala:70)
> at
> kafka.server.KafkaHealthcheck$SessionExpireListener.handleNewSession(KafkaHealthcheck.scala:104)
> at org.I0Itec.zkclient.ZkClient$6.run(ZkClient.java:735)
> at org.I0Itec.zkclient.ZkEventThread.run(ZkEventThread.java:71)
> {noformat}
--
This message was sent by Atlassian JIRA
(v6.3.15#6346)
