It seems you got OutOfMemoryError, which may leave the broker in a bad
state. You probably need a larger heap space.
Thanks,
Jun
On Fri, Apr 18, 2014 at 1:58 PM, Alex Demidko wrote:
> Last time saw this exception when tried to use rebalance leadership with
> kafka-preferred-replica-election.sh
Do you see any rebalances? The fetcher was stopped because it was shut
down, which typically happens during rebalances.
Thanks,
Jun
On Fri, Apr 18, 2014 at 11:08 AM, ankit tyagi wrote:
> I have checked that. There was no full gc at that time. I have attached
> jstat ouput too in my mail.
>
> I
Tried to reproduce this one more time. I was using kill -9 shutdown to test
reiliability, with graceful termination I haven't seen this problem to
arise. Leader node started complaining that ReplicaFetcherThread can't
connect to other node and that Producer can't send request to terminated
node, bu
Are you sure about that? Our latest tests show that loosing the drive in a
jbod setup makes the broker fail (unfortunately).
On Apr 18, 2014 9:01 PM, "Bello, Bob" wrote:
> Yes you would lose the topic/partitions on the drive. I'm not quite sure
> if Kafka can determine what topics/partitions are
When you are shutting down the restart node, did you see any warn/error on
the leader logs?
Guozhang
On Fri, Apr 18, 2014 at 1:58 PM, Alex Demidko wrote:
> Last time saw this exception when tried to use rebalance leadership with
> kafka-preferred-replica-election.sh. That's what got in logs:
>
Last time saw this exception when tried to use rebalance leadership with
kafka-preferred-replica-election.sh. That's what got in logs:
LeaderNode: just kafka.common.KafkaException: This operation cannot be
completed on a complete request without any other exceptions.
RestartedNode:
2014-04-18 2
Ok we will try that.
-Original Message-
From: Guozhang Wang [mailto:wangg...@gmail.com]
Sent: Friday, April 18, 2014 2:32 PM
To: users@kafka.apache.org
Subject: Re: KAFKA-717
KAFKA-1365 has been patched. Could you give it a try again after you have
tested it on dev environment?
On Fri
KAFKA-1365 has been patched. Could you give it a try again after you have
tested it on dev environment?
On Fri, Apr 18, 2014 at 1:15 PM, Seshadri, Balaji
wrote:
> The controller not failing over which I feel we got it resolved.
>
> The other fix is ZK node not getting deleted when preferred repl
Hello Alex,
I think this is a bug on the FetchResponseSend class. Just to confirm,
before the
kafka.common.KafkaException: This operation cannot be completed on a
complete request.
do you see other warn/error logs on the current leader?
Guozhang
On Fri, Apr 18, 2014 at 11:57 AM, Alexander Dem
The controller not failing over which I feel we got it resolved.
The other fix is ZK node not getting deleted when preferred replica election is
triggered.
https://issues.apache.org/jira/browse/KAFKA-1365
-Original Message-
From: Guozhang Wang [mailto:wangg...@gmail.com]
Sent: Friday,
Hi Balaji,
What issues do you have doing the upgrade?
On Fri, Apr 18, 2014 at 10:25 AM, Seshadri, Balaji wrote:
> Hi Jun,
>
> We could not move to 0.8.1 because of issues we have in upgrade.
>
> We are still in 0.8-beta1.
>
> Balaji
>
> -Original Message-
> From: Jun Rao [mailto:jun...
Yes you would lose the topic/partitions on the drive. I'm not quite sure if
Kafka can determine what topics/partitions are missing or not. I suggest you
try testing it.
- Bob
-Original Message-
From: Andrew Otto [mailto:ao...@wikimedia.org]
Sent: Friday, April 18, 2014 8:36 AM
To: use
Have tried to reproduce this error, and it occurs pretty consistently when
node being forcefully shutdown w/o graceful termination. When graceful
shutdown was successful no errors occur in a log when the instance was
rebooted starts.
On Fri, Apr 18, 2014 at 11:17 AM, Alex Demidko wrote:
> These
Success!
I can now failover without getting stuck into the logging loop. I am able to
failover between Kafka brokers. (Version 0.8.1)
I adjusted the following settings:
#(was 3)
controller.socket.timeout.ms=9
controlled.shutdown.enable=true
controlled.shutdown.max.retries=3
#(was 5000)
These on alive node:
2014-04-17 21:36:29,276 ERROR [ZkClient-EventThread-15] state.change.logger -
Controller 2 epoch 8 encountered error while electing leader for partition
[loadtest,143] due to: Preferred replica 1 for partition [loadtest,143] is
either not alive or not in the isr. Current le
I have checked that. There was no full gc at that time. I have attached
jstat ouput too in my mail.
I have concerned over why consumer fetcher thread are getting stopped.
On 18 Apr 2014 22:51, "Jun Rao" wrote:
> Have you looked at
>
> https://cwiki.apache.org/confluence/display/KAFKA/FAQ#FAQ-Why
Thanks Jun.
-Original Message-
From: Jun Rao [mailto:jun...@gmail.com]
Sent: Friday, April 18, 2014 11:37 AM
To: users@kafka.apache.org
Subject: Re: commitOffsets by partition 0.8-beta
We don't have the ability to commit offset at the partition level now. This
feature probably won't be
Any errors from the controller/state-change log?
Thanks,
Jun
On Fri, Apr 18, 2014 at 9:57 AM, Alex Demidko wrote:
> Hi,
>
> I’m performing a producing load test on two node kafka cluster built from
> the last 0.8.1 branch sources. I have topic loadtest with replication
> factor 2 and 256 parti
We don't have the ability to commit offset at the partition level now. This
feature probably won't be available until we are done with the consumer
rewrite, which is 3-4 months away.
If you want to do sth now and don't want to use SimpleConsumer, another
hacky way is to turn off auto offset commit
Thanks Jun! I added it to the ecosystem page.
On Fri, Apr 18, 2014 at 1:26 PM, Jun Rao wrote:
> Wow, Joe. That looks great. Could you add it to our wiki?
>
> Thanks,
>
> Jun
>
>
> On Fri, Apr 18, 2014 at 9:51 AM, Joe Stein wrote:
>
> > Hi, we started a new github project for using Kafka as the
Wow, Joe. That looks great. Could you add it to our wiki?
Thanks,
Jun
On Fri, Apr 18, 2014 at 9:51 AM, Joe Stein wrote:
> Hi, we started a new github project for using Kafka as the central point
> for all application and infrastructure metrics
> https://github.com/stealthly/metrics-kafka/.
>
Hi Jun,
We could not move to 0.8.1 because of issues we have in upgrade.
We are still in 0.8-beta1.
Balaji
-Original Message-
From: Jun Rao [mailto:jun...@gmail.com]
Sent: Friday, April 18, 2014 11:23 AM
To: users@kafka.apache.org
Subject: Re: KAFKA-717
Balaji,
0.8.0-BETA is too old
Balaji,
0.8.0-BETA is too old and we are not patching it any more. You probably can
try 0.8.0 or wait until 0.8.1.1 is out.
Thanks,
Jun
On Fri, Apr 18, 2014 at 8:26 AM, Seshadri, Balaji
wrote:
> I'm trying to apply the patch from KAFKA-717 for 0.8.0-BETA candidate and
> it fails.
>
> Error:
>
Have you looked at
https://cwiki.apache.org/confluence/display/KAFKA/FAQ#FAQ-Whyaretheremanyrebalancesinmyconsumerlog
?
Thanks,
Jun
On Fri, Apr 18, 2014 at 3:58 AM, ankit tyagi wrote:
> Hi,
>
> I am seeing consumer re-balances very frequently and getting socket
> reconnect exception. log is gi
Hi,
I’m performing a producing load test on two node kafka cluster built from the
last 0.8.1 branch sources. I have topic loadtest with replication factor 2 and
256 partitions. Initially both brokers are in ISR and leadership is balanced.
When in the middle of the load test one broker was resta
If you lose one drive in a JBOD setup you will just re-replicate the data
on that disk. It is similar to what you would do during RAID repair except
that instead of having the data coming 100% from the mirror drives the load
will be spread over the rest of the cluster.
The real downside of RAID is
Hi,
We have use case in DISH where we need to stop the consumer when we have issues
in proceeding further to database or another back end.
We update offset manually for each consumed message. There are 4 threads(e.g)
consuming from same connector and when one thread commits the offset there is
Hi, we started a new github project for using Kafka as the central point
for all application and infrastructure metrics
https://github.com/stealthly/metrics-kafka/.
We started off with implementing a Metrics Reporter (for Coda Hale's
metrics) which produces (reports) the statistics to a Kafka topi
I'm trying to apply the patch from KAFKA-717 for 0.8.0-BETA candidate and it
fails.
Error:
Patch failed:project/Build.scala
Project/Build.scala patch does not apply.
Please let me know if you guys have how to do it.
Thanks,
Balaji
> BOB> We are using RAID10. It was a requirement from our Unix guys. The
> rationale for this was we didn't want to lose just a disk and to have to
> rebuild/re-replicate 20TB of data. We haven't experienced any drive failures
> that I am aware of. We have had complete server failures, but the d
Hi,
I am seeing consumer re-balances very frequently and getting socket
reconnect exception. log is given below for more insights
[2014-04-18
16:02:52.061][kafka.coms.consumer.kafka_topic_coms_esb_prod_coms.coms-timemachine.coms.coms04.snapdeal.com_coms04.snapdeal.com-1397812122323-509d9663_watc
31 matches
Mail list logo