RE: ConsumerRebalanceFailedException

25, 2014 4:03 PM To: users@kafka.apache.org Subject: Re: ConsumerRebalanceFailedException Could you send around the consumer log when it throws ConsumerRebalanceFailedException. It should state the reason for the failed rebalance attempts. Thanks, Neha On Tue, Feb 25, 2014 at 12:01 PM, Yu, Li

ConsumerRebalanceFailedException

Hi all, I tried to reproduce this exception. In case one, when no broker was running, I launched all consumers and got this exception. In case two, while the consumers and brokers were running, I shutdown all brokers one by one and did not see this exception. I wonder why in case two this except

RE: broker offline

-register upon receiving the session timeout. You can re-produce this issue by signal pause the ZK process. Guozhang On Fri, Feb 14, 2014 at 12:15 PM, Yu, Libo wrote: > Hi team, > > We have three brokers on our production cluster. I noticed two of them > somehow got offline and then r

broker offline

Hi team, We have three brokers on our production cluster. I noticed two of them somehow got offline and then re-registered with zookeeper and got back online. It seems the issue was caused by some zookeeper issue. So I want to know what may be the possible cases of the issue. If I want to reprod

RE: some brokers cannot register themselves with zookeeper

version are you using? Pre-0.8.1 there is a bug that can cause a registration path to be deleted: https://issues.apache.org/jira/browse/KAFKA-992 And this has been fixed in 0.8.1 Guozhang On Tue, Feb 11, 2014 at 1:16 PM, Yu, Libo wrote: > Hi team, > > This is an issue that has frus

some brokers cannot register themselves with zookeeper

Hi team, This is an issue that has frustrated me for quit some time. One of our clusters has three hosts. In my startup script, three zookeeper processes are brought up first followed by three kafka processes. The problem we have is that after three kafka processes are up, only one broker has b

RE: How to compile with a newer version of zookeeper

When I telnet to the zookeeper and type "status", this is what I got: Zookeeper version: 3.3.3-1203054, built on 11/17/2011 05:47 GMT Is that 3.3.4? So 0.8 final also uses 3.3.4, is that right? Thanks. Regards, Libo -Original Message- From: Neha Narkhede [mailto:neha.narkh...@gmail.com

RE: Consumer's behavior when brokers are temporarily not available.

lable. Yes bouncing the process will allow you to consume again. Also would you mind giving 0.8 final a try? It is much more stable compared to 0.8 beta. Thanks, Neha On Feb 7, 2014 6:49 AM, "Yu, Libo" wrote: > We are using 0.8 beta1. Our zookeeper had some issue which in turn &g

RE: Consumer's behavior when brokers are temporarily not available.

ll trigger rebalances. Thanks, Jun On Thu, Feb 6, 2014 at 10:46 AM, Yu, Libo wrote: > While the broker is not available (caused by zookeeper issue), the > rebalance will fail. Should rebalance succeed in this case? Thanks. > > > Regards, > > Libo > > > -Ori

RE: Consumer's behavior when brokers are temporarily not available.

ns out of the retries, it needs to be restarted to consume again. On Thu, Feb 6, 2014 at 9:05 AM, Yu, Libo wrote: > Hi folks, > > This is what we experienced recently: > Some zookeeper's issue made broker unavailable for a short period of time. > On the consumer side, thi

RE: Consumer's behavior when brokers are temporarily not available.

gt; > > > > On Thu, Feb 6, 2014 at 9:05 AM, Yu, Libo wrote: > > > Hi folks, > > > > This is what we experienced recently: > > Some zookeeper's issue made broker unavailable for a short period of > time. > > On the consumer side, this triggere

Consumer's behavior when brokers are temporarily not available.

Hi folks, This is what we experienced recently: Some zookeeper's issue made broker unavailable for a short period of time. On the consumer side, this triggered rebalance and rebalanced failed after four tries. So while should we expect while the broker is not up? Should consumer keep trying to reb

num.partitions

Hi team, I believe num.partitions is for automatic topic creation. Is that right? The default number of partition for kafka-create-topic.sh is 1. So Will num.partitions impact kafka-create-topic.sh? Thanks. Regards, Libo

RE: purging a topic

You would need to first stop the consumer, update the offset in ZK and then restart the consumer. Also, have you looked at the tool ImportZkOffsets? Thanks, Jun On Tue, Jan 14, 2014 at 12:38 PM, Yu, Libo wrote: > Hi folks, > > I am writing a tool to "purge" the pending topics

purging a topic

Hi folks, I am writing a tool to "purge" the pending topics for a user. Assume the user has never consumed this topic previously. If I create all the nodes on the path /consumers/[myuser]/offsets/[mytopic]/[partition] and put the maximum available offset to the node, is that enough to let the con

RE: ConsumerRebalanceFailedException

Hi Jun, zookeeper.session.timeout.ms is used in a broker's configuration and manages brokers' registration with zk. Does it apply to consumer as well? Thanks. Regards, Libo -Original Message- From: Jun Rao [mailto:jun...@gmail.com] Sent: Monday, December 30, 2013 11:13 AM To: users@

RE: upgrade from beta1 to 0.81

8.1 is working in stable at LinkedIn now. > > Guozhang > > > On Thu, Dec 19, 2013 at 10:52 AM, Yu, Libo wrote: > > > I also want to know how stable the 0.81 will be, compared with 0.8 > > or 0.8-beta1. > > > > Regards, > > > > Libo > &g

RE: upgrade from beta1 to 0.81

is 0.8.1, will there be a 'release' of this soon, or are there still significant open issues? Thanks, Jason On Thu, Dec 19, 2013 at 12:17 PM, Guozhang Wang wrote: > Libo, yes the upgrade from 0.8 to 0.8.1 can be done in place. > > Guozhang > > > On Thu, Dec 1

RE: upgrade from beta1 to 0.81

beta1 to 0.81 Libo, yes the upgrade from 0.8 to 0.8.1 can be done in place. Guozhang On Thu, Dec 19, 2013 at 8:57 AM, Yu, Libo wrote: > Hi folks, > > As the tools in 0.8 are not stable and we don't want to take the risk. > we want to skip 0.8 and upgrade from beta1 t

upgrade from beta1 to 0.81

Hi folks, As the tools in 0.8 are not stable and we don't want to take the risk. we want to skip 0.8 and upgrade from beta1 to 0.81 directly. So my question is whether we can do an in place upgrade and let 0.81 use beta1's zk and kf data. Assume that we will disable log compaction. Thanks. Regard

RE: a consumer question

will return false instead of > throwing > > an exception. > > > > Guozhang > > > > > > On Tue, Dec 17, 2013 at 11:53 AM, Yu, Libo wrote: > > > > > Sorry, a typo. Correct my question. When consumer.timeout.ms is > > > set to > > 0, &

RE: a consumer question

Jun On Tue, Dec 17, 2013 at 4:57 PM, Guozhang Wang wrote: > If there is no more messages, hasNext will return false instead of > throwing an exception. > > Guozhang > > > On Tue, Dec 17, 2013 at 11:53 AM, Yu, Libo wrote: > > > Sorry, a typo. Correct my questi

RE: a consumer question

: Tuesday, December 17, 2013 12:40 AM To: users@kafka.apache.org Subject: Re: a consumer question If there is a message, hasNext() returns true, not throwing an exception. Thanks, Jun On Mon, Dec 16, 2013 at 11:29 AM, Yu, Libo wrote: > Hi folks, > > For this parameters, if consumer.t

a consumer question

Hi folks, For this parameters, if consumer.timeout.ms is set to 0, whenever I call ConsumerIterator's hasNext(), if there is a message available, a timeout exception will be thrown. Is my understanding correct? Thanks. consumer.timeout.ms -1 Throw a timeout exception to the consumer if no mes

RE: cluster expansion

venly distributed on all six brokers. If I use reassignment tool in 0.81 with 0.8 broker, will that work and get around the bugs? Your broker also needs to be on 0.8.1 for it to work correctly. On Mon, Dec 16, 2013 at 9:06 AM, Yu, Libo wrote: > If we have six brokers, and a topic has three pa

RE: cluster expansion

-Original Message- From: Neha Narkhede [mailto:neha.narkh...@gmail.com] Sent: Monday, December 16, 2013 10:26 AM To: users@kafka.apache.org Subject: RE: cluster expansion They will be evenly distributed across the nodes in the cluster. Thanks, Neha On Dec 16, 2013 6:42 AM, "Yu,

RE: cluster expansion

will remain on the original brokers. > > You > > > could either reassign some partitions from all topics to the new > brokers > > or > > > you could add partitions to the new brokers for each topic. In > > > 0.8.0 > > there > > > is now

cluster expansion

Hi folks, There are three brokers running 0.8-beta1 in our cluster currently. Assume all the topics have six partitions. I am going to add another three brokers to the cluster and upgrade all of them to 0.8. My question is after the cluster is up, will the partition be evenly distributed to all

RE: error from adding a partition

AM To: users@kafka.apache.org Subject: Re: error from adding a partition Tried this on the 0.8.0 release and it works for me. Could you make sure there are no duplicated kafka jars? Thanks, Jun On Tue, Dec 10, 2013 at 7:08 AM, Yu, Libo wrote: > Hi folks, > > I got this error when I tri

error from adding a partition

Hi folks, I got this error when I tried to test the partition addition tool. bin/kafka-add-partitions.sh --partition 1 --topic libotesttopic --zookeeper xx.xxx.xxx.xx: adding partitions failed because of kafka.admin.AdminUtils$.assignReplicasToBrokers(Lscala/collection/Seq;)Lscala/collec

RE: set new retention size on the fly

+release+plan Thanks, Jun On Wed, Dec 4, 2013 at 8:15 AM, Yu, Libo wrote: > Thanks for the clarification. I am just curious about how this works out. > If we can change the retention size with "kafka-topics.sh --alter", > will the new retention size be updated to the serve

RE: set new retention size on the fly

g tool. 0.8.1 makes all per-topic > configuration dynamic and updatable via a command line tool. > > -Jay > > > On Tue, Dec 3, 2013 at 1:23 PM, Yu, Libo wrote: > > > Hi Neha, > > > > "0.8.1 includes the ability to dynamically change per top

RE: set new retention size on the fly

t 10:21 AM, Yu, Libo wrote: > Hi folks, > > For 0.8, it is possible to add a partition dynamically. Is it possible > to increase the retention size on the fly? This feature will be very > useful for operation. I know rolling start can pick up the change but > it takes too much effort. Thanks. > > Libo > >

set new retention size on the fly

Hi folks, For 0.8, it is possible to add a partition dynamically. Is it possible to increase the retention size on the fly? This feature will be very useful for operation. I know rolling start can pick up the change but it takes too much effort. Thanks. Libo

RE: ConsumerRebalanceFailedException

, Jun On Mon, Dec 2, 2013 at 6:57 AM, Yu, Libo wrote: > Actually, I saw this line in the log : can't rebalance after 4 retries. > What should I expect in this case? All consumers threads failed or > only some of them? > If I increase the number of retries or delay between retr

RE: upgrade from 0.8-beta1 to 0.8

, Nov 29, 2013 at 6:35 AM, Yu, Libo wrote: > Hi team, > > Currently we are using 0.8-beta1. We plan to upgrade to 0.8. My > concern is whether we need to purge all existing kafka and zookeeper > data on the hard drive for this upgrade. In other words, can 0.8 use > 0.8-beta1 k

RE: ConsumerRebalanceFailedException

ill not be consumed by any consumers. Thanks, Jun On Fri, Nov 29, 2013 at 10:44 AM, Yu, Libo wrote: > You are right, Joe. I checked our brokers' log. We have three brokers. > All of them failed to connect to zk at some point. > So they were offline and later reregistered themselves

RE: ConsumerRebalanceFailedException

http://www.twitter.com/allthingshadoop> ****/ On Fri, Nov 29, 2013 at 11:31 AM, Yu, Libo wrote: > We found our consumer stopped working after this exception occurred. > Can the consumer recover from such an exception? > > Regards, > > Libo > > > -O

RE: ConsumerRebalanceFailedException

We found our consumer stopped working after this exception occurred. Can the consumer recover from such an exception? Regards, Libo -Original Message- From: Florin Trofin [mailto:ftro...@adobe.com] Sent: Tuesday, July 16, 2013 4:20 PM To: users@kafka.apache.org Subject: Re: ConsumerReb

upgrade from 0.8-beta1 to 0.8

Hi team, Currently we are using 0.8-beta1. We plan to upgrade to 0.8. My concern is whether we need to purge all existing kafka and zookeeper data on the hard drive for this upgrade. In other words, can 0.8 use 0.8-beta1 kafka and zookeeper data on the hard drive? Thanks. Regards, Libo

RE: addition of a partition

, 2013 at 12:28 PM, Yu, Libo wrote: > Hi team, > > I am reading this link: > > https://cwiki.apache.org/confluence/display/KAFKA/Replication+tools#Re > plicationtools-5.AddPartitionTool and this JIRA > https://issues.apache.org/jira/i#browse/KAFKA-1030. I have a couple of &

RE: about compiling broker in 0.8 branch

Source Security LLC http://www.stealth.ly Twitter: @allthingshadoop <http://www.twitter.com/allthingshadoop> / On Thu, Nov 28, 2013 at 4:24 PM, Yu, Libo wrote: > Hi team, > > For the current 0.8 branch, is it recommended to compil

about compiling broker in 0.8 branch

Hi team, For the current 0.8 branch, is it recommended to compile it with Scala 2.10? I remember someone said previously it is best to compile the broker with Scala 2.80. Thanks. Regards, Libo

addition of a partition

Hi team, I am reading this link: https://cwiki.apache.org/confluence/display/KAFKA/Replication+tools#Replicationtools-5.AddPartitionTool and this JIRA https://issues.apache.org/jira/i#browse/KAFKA-1030. I have a couple of questions. After adding a partition by using the tool, should the consumer

RE: retention size

I did a restart and the issue was gone. It could be that we changed the retention size and did not restart the brokers to pick up the change. Thanks for your help. Regards, Libo -Original Message- From: Yu, Libo [ICG-IT] Sent: Friday, November 22, 2013 10:53 AM To: '

RE: retention size

producer sticks to a partition for the metadata refresh period - so if your test run isn't long enough some partitions may be more loaded than the others. On Thu, Nov 21, 2013 at 06:28:39PM +0000, Yu, Libo wrote: > Hi team, > > We have 3 brokers in a cluster. The replication factor i

retention size

Hi team, We have 3 brokers in a cluster. The replication factor is 2. I set the default retention size to 3G bytes. I published 12G data to a topic, which is enough to fully load all partitions. I assume on each broker the partition size should be 3G. However, it is only 1.4G for one partation.

NodeExists Keeper exception

Hi team, Still this is from beta1. I notice this exception occurred frequently in our broker logs. [2013-11-14 21:09:58,714] INFO Got user-level KeeperException when proce ssing sessionid:0x24250e816b000df type:create cxid:0x26b zxid:0x fffe txntype:unknown reqpath:n/a Error Path:/con

RE: broker exception

...@gmail.com] Sent: Thursday, November 14, 2013 12:27 PM To: users@kafka.apache.org Subject: Re: broker exception Are you using ZK 3.3.4? This seems to be caused by a bug in 3.3.3 and 3.3.0. https://issues.apache.org/jira/browse/ZOOKEEPER-1115 Thanks, Jun On Thu, Nov 14, 2013 at 5:25 AM, Yu, Libo

RE: Leader not local

r log? Guozhang On Thu, Nov 14, 2013 at 8:23 AM, Yu, Libo wrote: > Hi team, > > I saw this line within a long time span in our logs for the same topic > and partition. > [2013-11-14 11:13:41,647] WARN [KafkaApi-1] Produce request with > correlation id 529240 from client

Leader not local

Hi team, I saw this line within a long time span in our logs for the same topic and partition. [2013-11-14 11:13:41,647] WARN [KafkaApi-1] Produce request with correlation id 529240 from client on partition [mytopic,2] failed due to Leader not local for partition [mytopic,2] on broker 1 (kafka

will this cause message loss?

Hi team, We are using beta1. I am going to delete all topics and create them with more partitions. But I don't want to lose any messages. Assume the consumers are online all the time for the following steps. The consumer's auto.offset.reset is set to largest. 1 stop publishing to the brokers.

broker exception

Hi team, This exception occurs regularly on our brokers. When it occurs, a broker will lose its leader role but still in ISR. And running preferred-leader-election script may rebalance the leadership but in some cases it does not help. [2013-11-14 08:04:40,001] INFO Processed session terminatio

RE: Purgatory

I read it and tried to understand it. It would be great to add a summary at the beginning about what it is and how it may impact a user. Regards, Libo -Original Message- From: Joel Koshy [mailto:jjkosh...@gmail.com] Sent: Friday, November 08, 2013 2:01 AM To: users@kafka.apache.org Sub

RE: add partition tool in 0.8

Thanks for your reply, Joel. Regards, Libo -Original Message- From: Joel Koshy [mailto:jjkosh...@gmail.com] Sent: Thursday, November 07, 2013 5:00 PM To: users@kafka.apache.org Subject: Re: add partition tool in 0.8 > > kafka-add-partitions.sh is in 0.8 but not in 0.8-beta1. Therefor

add partition tool in 0.8

Hi team, Here is what I want to do: We are using 0.8-beta1 currently. We already have some topics and want to add partitions for them. kafka-add-partitions.sh is in 0.8 but not in 0.8-beta1. Therefore we cannot use this tool with 0.8-beta1. If I download latest 0.8 and compile it, can I use its

RE: question about default key

Got it. Thanks. Regards, Libo -Original Message- From: Neha Narkhede [mailto:neha.narkh...@gmail.com] Sent: Thursday, October 24, 2013 10:09 AM To: users@kafka.apache.org Subject: Re: question about default key The default key is null. Thanks, Neha On Oct 24, 2013 6:47 AM, "Yu,

question about default key

Hi team, If I don't specify a key when publishing a message, a default key will be generated. In this case, how long is the default key and will the consumer get this default key? Thanks. Libo

KeyedMessage question

Hi team, For the message type in KeyedMessage, I can use String or byte[]. Is there any difference in terms of the actual data transferred? Regards, Libo

key partitioner

Hi team, According to the document, the default partitioner hashes the key string and assign the message to a broker. Could you give a brief introduction to the hash algorithm? If a long timestamp (in hex format) is used as key, will the messages be distributed evenly to all partitions? Assume

RE: producer API thread safety

, 2013 at 9:14 AM, Yu, Libo wrote: > Hi team, > > Is it possible to use a single producer with more than one threads? I > am not sure If its send() is thread safe. > > Regards, > > Libo > > -- -- Guozhang

producer API thread safety

Hi team, Is it possible to use a single producer with more than one threads? I am not sure If its send() is thread safe. Regards, Libo

RE: is it possible to commit offsets on a per stream basis?

ssage 'markedForCommit', and not the last 'consumed' offset, which may or may not have succeeded. This way, consumer code can just call 'markForCommit()' after successfully processing each message successfully. Does that make sense? On Mon, Sep 9, 2013 at 5:21 P

RE: Strategies for improving Consumer throughput

You can use a thread pool to write to hbase. And create another pool of consumer threads. Or add more consumer processes. The bottleneck is writing to Hbase in this case. Regards, Libo -Original Message- From: Graeme Wallace [mailto:graeme.wall...@farecompare.com] Sent: Wednesday, Oct

bandwidth usage issue

Hi team, Here is a usage case: Assume each host in a kafka cluster a gigabit network adaptor. And the incoming traffic is 0.8gbps and at one point all the traffic goes to one host. The remaining bandwidth is not enough for the followers to replicate messages from this leader. To make sure no b

RE: [jira] [Updated] (KAFKA-1046) Added support for Scala 2.10 builds while maintaining compatibility with 2.8.x

Hi team, Is it safe to apply the 0.8 patch to 0.8 beta1? Regards, Libo -Original Message- From: Joe Stein [mailto:crypt...@gmail.com] Sent: Friday, September 13, 2013 4:10 PM To: d...@kafka.apache.org; users@kafka.apache.org Subject: Re: [jira] [Updated] (KAFKA-1046) Added support for

RE: [jira] [Updated] (KAFKA-1046) Added support for Scala 2.10 builds while maintaining compatibility with 2.8.x

Answer my own question: When I tried to apply the patch to 0.8 beta1, I got many errors and had to skip it. Regards, Libo -Original Message- From: Yu, Libo [ICG-IT] Sent: Tuesday, September 17, 2013 3:33 PM To: 'users@kafka.apache.org'; 'd...@kafka.apache.org'

RE: monitoring followers' lag

, September 10, 2013 8:48 PM To: users@kafka.apache.org Subject: Re: monitoring followers' lag It should be "kafka.server":type="ReplicaFetcherManager",name="Replica-MaxLag" - can you confirm and mind updating the wiki if this is the case? Thanks, Joel On Tue,

monitoring followers' lag

Hi team, I wonder if anybody can give detailed instructions on how to monitor the followers' lag by using JMX. Thanks. Regards, Libo

RE: monitoring followers' lag

mailto:jun...@gmail.com] Sent: Tuesday, September 10, 2013 11:01 AM To: users@kafka.apache.org Subject: Re: monitoring followers' lag Have you looked at the updated docs in http://kafka.apache.org/documentation.html#monitoring ? Thanks, Jun On Tue, Sep 10, 2013 at 7:59 AM, Yu, Libo wrote: >

RE: implicit default minimum retention size per partition is 4GB.

o messages translation yourself. As for setting replica.lag.max.messages, you can observe the max lag in the follower and set replica.lag.max.messages to be a bit larger than that. I am curious to know the observed max lag in your use case. Thanks, Jun On Tue, Sep 10, 2013 at 6:46 AM, Yu, L

RE: monitoring followers' lag

e.org/documentation.html#monitoring ? Thanks, Jun On Tue, Sep 10, 2013 at 7:59 AM, Yu, Libo wrote: > Hi team, > > I wonder if anybody can give detailed instructions on how to monitor > the followers' lag by using JMX. Thanks. > > Regards, > > Libo > >

implicit default minimum retention size per partition is 4GB.

Hi team, For default broker configuration, replica.lag.max.messages is 4000 and message.max.bytes is 1Mb. In the extreme case, the follower(s) could lag by 4000 messages. The leader must save at least 4000 messages to allow follower(s) to catch up. So the minimum retention size is 4000Mb=4Gb. I

RE: is it possible to commit offsets on a per stream basis?

. Thanks, Neha On Mon, Sep 9, 2013 at 9:08 AM, Yu, Libo wrote: > If one connector is used for a single stream, when there are many > topics/streams, will that cause any performance issue, e.g. too many > connections or too much memory or big latency? > > Regards, > > Lib

RE: is it possible to commit offsets on a per stream basis?

ocessed (seems a change to the connector itself > > > might expose a way to > > use > > > auto offset commit, and have it never commit a message until it is > > > processed). But that would be a change to the > > > ZookeeperConsumerConnectorEssenti

RE: is it possible to commit offsets on a per stream basis?

the same time? This is a better approach as there is no complex locking involved. Thanks, Neha On Thu, Aug 29, 2013 at 10:28 AM, Yu, Libo wrote: > Hi team, > > This is our current use case: > Assume there is a topic with multiple partitions. > 1 Create a connector first and

is it possible to commit offsets on a per stream basis?

Hi team, This is our current use case: Assume there is a topic with multiple partitions. 1 Create a connector first and create multiple streams from the connector for a topic. 2 Create multiple threads, one for each stream. You can assume the thread's job is to save the message into the database

RE: zookeeper session time out

n Wed, Aug 28, 2013 at 1:09 PM, Yu, Libo wrote: > Hi team, > > We notice when the incoming throughput is very high, the broker has to > delete old log files to free up disk space. That caused some kind of > blocking > (latency) and > frequently the broker's zookeeper s

RE: questtion about log.retention.bytes

nted in > > http://kafka.apache.org/documentation.html#brokerconfigs > > > > "Note that all per topic configuration properties below have the > > format > of > > csv (e.g., "topic1:value1,topic2:value2")." > > Thanks, > > Jun > >

zookeeper session time out

Hi team, We notice when the incoming throughput is very high, the broker has to delete old log files to free up disk space. That caused some kind of blocking (latency) and frequently the broker's zookeeper session times out. Currently our zookeeper time out threshold is 4s. We can increase it. Bu

RE: questtion about log.retention.bytes

Hi Jun, In a previous email thread http://markmail.org/search/?q=kafka+log.retention.bytes#query:kafka%20log.retention.bytes+page:1+mid:qnt4pbq47goii2ui+state:results, you said log.retention.bytes is for each partition. Could you clarify on that? Say if I have a topic with three partitions. I wa

RE: questions about ISR

, having a large replica.lag.time.max.ms may delay the committing of a message. Thanks, Jun On Tue, Aug 27, 2013 at 6:37 AM, Yu, Libo wrote: > Thanks, Jun. That is very helpful. However, I still have a couple of > questions. "We have a min fetch rate JMX in the broker". How t

Question about preferred replica leader election tool

Hi, We have three brokers in our kafka cluster. For all topics, the replica factor is two. Here is the distribution of leaders. After I ran the leader election tool, nothing happened. In this list, the first broker in ISR is the leader. I assume after running the tool, the first broker is repli

RE: questions about ISR

> That's right. You shouldn't need to restart the whole cluster for a > broker > > to rejoin ISR. Do you see many ZK session expirations in the brokers > > (search for "(Expired)"? If so, you may need to tune the GC on the > broker. > > > > T

RE: questions about ISR

ISR When a broker is restarted, it will automatically catch up from the leader and will join ISR when it's caught up. Are you not seeing this happening? Thanks, Jun On Fri, Aug 23, 2013 at 11:33 AM, Yu, Libo wrote: > Hi, > > When a broker is not in a topic's ISR, will it

RE: delete a topic

-729 Thanks, Neha On Fri, Aug 23, 2013 at 10:52 AM, Yu, Libo wrote: > I will give it a try. I know how to delete log files. But to delete > the zookeeper data, do I only need to run the delete script? > > Regards, > > Libo > > > -Original Message

questions about ISR

Hi, When a broker is not in a topic's ISR, will it try to catch up to go back to ISR itself? Or do we have to restart it? We can increase replica.lag.time.max.ms and replica.lag.max.messages to let brokers stay longer in ISR. Is that good practice? Still this is related to the first questions. W

RE: delete a topic

to bounce the entire cluster once you've deleted the zookeeper and kafka data for the topic in question. Can you give it a try and let us know how it went? Thanks, Neha On Fri, Aug 23, 2013 at 10:15 AM, Yu, Libo wrote: > Hi Neha, > > One more questions. Assume I want to dele

RE: KeeperErrorCode = BadVersion

shrinking ISR or electing a new leader for the same partition. Could you please file a JIRA to improve the quality of logging in this case? Thanks, Neha On Fri, Aug 23, 2013 at 10:28 AM, Yu, Libo wrote: > Hi team, > > During normal operation, all of a sudden, we found many exceptions

KeeperErrorCode = BadVersion

Hi team, During normal operation, all of a sudden, we found many exceptions in the log like this: It seems one thread' zookeeper's data is written unexpectedly by some other thread. Any expertise will be appreciated. [2013-08-23 13:17:00,622] INFO Partition [our.own.topic one.default,0] on bro

RE: delete a topic

-1021 Thanks, Neha On Fri, Aug 23, 2013 at 6:34 AM, Yu, Libo wrote: > Hi Neha, > > "Wipe out the cluster" Do you mean you uninstall the cluster and > reinstall it? > Or you just delete all kafka data and zookeeper data for the cluster? > This is not a blocking issu

RE: Producer message ordering problem

An auto-increment index can be assigned to a message as a key when it is being published. The consumer can monitor this index when receiving. If the expected message does not show up, buffer all received messages in a hashtable (use index as hash key) until it is received. Then handle all mes

consumer question

Hi team, Right now, from a stream, an iterator can be obtained which has a blocking hasNext(). So what is the implementation behind the iterator? I assume there must be queue and the iterator monitors the queue. And a separate thread fetches data and feeds to the queue when it is almost empty.

RE: delete a topic

Right now it appears to work but > > >> > doesn't > which > > >> is > > >> > clearly not good. > > >> > > > >> > -Jay > > >> > > > >> > > > >> > On Thu, Aug 22, 2013 at 10:57 AM, Neha N

partition does not exist (but it is there).

This is from the broker 3's log: [2013-08-22 15:40:02,984] WARN [KafkaApi-3] Fetch request: Partition [tes t.replica1.default,0] doesn't exist on 3 (kafka.server.KafkaApis) Here is what list topic command shows: topic: test.replica1.defaultpartition: 0leader: 3 replicas: 3 isr:

RE: delete a topic

feature in Kafka 0.8. So > > any manual attempts to do so might have a negative impact on functionality. > > > > Thanks, > > Neha > > > > > > On Thu, Aug 22, 2013 at 10:30 AM, Yu, Libo wrote: > > > > > Hi team, > > > >

RE: issue with kafka-preferred-replica-election.sh

: https://issues.apache.org/jira/browse/KAFKA-1019 Guozhang On Wed, Aug 21, 2013 at 11:27 AM, Yu, Libo wrote: > We never deleted it. Either it was never created or deleted somehow. > > Regards, > > Libo > > > -Original Message- > From: Guozhang Wang [mailto

RE: delete a topic

ts to do so might have a negative impact on functionality. Thanks, Neha On Thu, Aug 22, 2013 at 10:30 AM, Yu, Libo wrote: > Hi team, > > When I delete a topic, the topic is deleted from zookeeper but its log > files are not deleted from Brokers. > > When I restart a broker,

delete a topic

Hi team, When I delete a topic, the topic is deleted from zookeeper but its log files are not deleted from Brokers. When I restart a broker, the broker will try to sync the log files whose topic has been deleted. Manually deleting the log files will resolve the issue. Should broker ignore log

RE: ordering

] Sent: Thursday, August 22, 2013 12:01 AM To: users@kafka.apache.org Subject: Re: ordering Actually, I am not sure if I understand the trouble that you mentioned. Could you elaborate that a bit more? Thanks, Jun On Wed, Aug 21, 2013 at 12:30 PM, Yu, Libo wrote: > Hi, > > This is f

  1   2   >