Thanks Jun Phil!
Shone
On Thu, Jun 13, 2013 at 12:00 AM, Jun Rao jun...@gmail.com wrote:
Yes, we just have customized encoder that encodes the first 4 bytes of md5
of the schema, followed by Avro bytes.
Thanks,
Jun
On Wed, Jun 12, 2013 at 9:50 AM, Shone Sadler shone.sad...@gmail.com
Hi Jun,
I was using the 0.8 branch with 2 commits behind but now I am using the
latest with the same issue. 3 topics A,B,C, created automatically with
replication factor of 2 and partitions 2. 2 brokers (0 and 1).
List of topics in zookeeper is the following:
topic: A partition: 0leader: 1
Looking at Jun’s ApacheCon slides (
http://www.slideshare.net/junrao/kafka-replication-apachecon2013) slide 21
titled, ‘Data Flow in Replication’ there are three possible durability
configurations which tradeoff latency for greater persistence guarantees.
The third row is the ‘no data loss’
I've tried the console producer, so I will assume that's not related with
the producer. I keep seeing the same entries in the producer from time to
time:
[2013-06-13 11:04:00,670] WARN Error while fetching metadata
[{TopicMetadata for topic C -
No partition metadata for topic C due to
Jun - thanks again. This is very helpful.
Philip
On Jun 12, 2013, at 9:50 PM, Jun Rao jun...@gmail.com wrote:
Actually, you are right. This can happen on a single topic too, if you have
more than one consumer thread. Each consumer thread pulls data from a
blocking queue, one or more
I think I know what's happening:
I tried to run both brokers and ZK on the same machine and it worked. I
also attempted to do the same but with a ZK node on other machine and it
also worked.
My guess is something related with ports. All the machines are on EC2 and
there might be something
No. It only means that messages are written to all replicas in memory. Data
is flushed to disk asynchronously.
Thanks,
Neha
On Jun 13, 2013 3:29 AM, Jonathan Hodges hodg...@gmail.com wrote:
Looking at Jun’s ApacheCon slides (
http://www.slideshare.net/junrao/kafka-replication-apachecon2013)
Have you looked at #3 in http://kafka.apache.org/faq.html?
Thanks,
Jun
On Thu, Jun 13, 2013 at 6:41 AM, Alexandre Rodrigues
alexan...@blismedia.com wrote:
I think I know what's happening:
I tried to run both brokers and ZK on the same machine and it worked. I
also attempted to do the
Hi all,
First, thanks to Tim (from Rabbit) and Jonathan for moving this thread
along. Jonathan, I hope you found my links to the data model docs,
and Tim's replies, helpful.
Has everyone got what they wanted from this thread?
alexis
On Tue, Jun 11, 2013 at 5:49 PM, Jonathan Hodges
I have but this is a different thing. It's related with ports and security
groups and not with the bind addresses. It's solved now.
Thanks
On 13 June 2013 15:42, Jun Rao jun...@gmail.com wrote:
Have you looked at #3 in http://kafka.apache.org/faq.html?
Thanks,
Jun
On Thu, Jun 13, 2013
Hi all, my team is proposing a novel
way of using Kafka and I am hoping someone can help do a sanity check on this:
1. When a user logs
into our website, we will create a “logged in” event message in Kafka
containing the user id.
2. 30+ systems
(consumers each in their own consumer groups)
Hi Alexis,
This was very helpful and I also appreciate both yours and Tim's input
here. It clears up the cases for when to use Rabbit or Kafka. What is
great is they are both open source with vibrant communities behind them.
-Jonathan
Go
On Jun 13, 2013 8:45 AM, Alexis Richardson
Hi,
In my application, I am storing user events, and I want to partition the
storage by day.
So at the end of a day, I want to take that file and ship it to s3 or
another server as a backup. This way I can replay the events for a
specific day if needed.
These events also have to be in order.
Hi Josh,
The idea looks very interesting. I just had one doubt.
1. A user logs in. His login id is sent on a topic
2. Other systems (consumers on this topic) consumer this message and
publish their results to another topic
This will be happening without any particular order for hundreds of
As an update, this continues to affect us.
First I'd like to note ways in which my issues seems different than
KAFKA-278,
* I did not add a new broker or a new topic, this topic has been in use on
two existing brokers for months
* The topic definitely exists on both brokers. The topic/data
Hi Mahendra, I think that is where it gets a little tricky. I think it would
work something like this:
1. Web sends login event for user user123 to topic GUEST_EVENT.
2. All of the systems consume those messages and publish the data messages to
topic GUEST_DATA.user123.
3. The
Also since you're going to be creating a topic per user, the number of
concurrent users will also be a concern to Kafka as it doesn't like massive
amounts of topics.
Tim
On Thu, Jun 13, 2013 at 10:47 AM, Josh Foure user...@yahoo.com wrote:
Hi Mahendra, I think that is where it gets a little
Ah yes, I had read that Kafka likes under 1,000 topics but I wasn't sure if
that was really a limitation. In principle I wouldn't mind having all guest
events placed on the GUEST_DATA queue but I thought that by having more
topics I could minimize having consumers read messages only to discard
You know what, it's likely this is all because I'm running a bad fork of
Kafka 0.7.2 for Scala 2.10 (on the producers/consumers) since that's the
version we've standardized on.
Behavior in 2.9.2 with the official Kafka 0.7.2 release seems much more
normal -- I'm working on downgrading all our
Spot on. This one was of the areas that we had to workaround. Remember
that there is a 1:1 relationship of topics to directories and most file
systems don't like 10s of thousands of directories. We found on practice
that 60k per machine was a practical limit using I believe EXT3FS
On
Hello -- is it possible for our code to stall a ConsumerConnector from
doing any consuming for, say, 30 seconds, until we can be sure that
all other ConsumeConnectors are rebalanced?
It seems that the first ConsumerConnector to come up is prefetching
some data, and we end up with duplicate
Just to be clear, I'm not asking that we solve duplicate messages on
crash before commit to Zookeeper, just an apparent problem where if
Kafka has some data, and we start on ConsumerConnectors, we get dupe
data since some Consumers come up before others.
Any help?
Philip
On Thu, Jun 13, 2013 at
Thanks. Which version of Intellij are you using?
Jun
On Thu, Jun 13, 2013 at 10:20 AM, Dragos Manolescu
dragos.manole...@servicenow.com wrote:
Hmm, I've just pulled 0.8.0-beta1-candidate1, removed .idea* from my
top-level directory, executed gen-idea, and then opened and built the
project
Are you messages compressed in batches? If so, some dups are expected
during rebalance. In 0.8, such dups are eliminated. Other than that,
rebalance shouldn't cause dups since we commit consumed offsets to ZK
before doing a rebalance.
Thanks,
Jun
On Thu, Jun 13, 2013 at 7:34 PM, Philip O'Toole
Jun -- thanks. We're using 0.72.
No, the messages are not compressed, and since we do appear to be
seeing dupes in our tests, it indicates our own code is buggy.
Thanks,
Philip
On Thu, Jun 13, 2013 at 9:15 PM, Jun Rao jun...@gmail.com wrote:
Are you messages compressed in batches? If so, some
25 matches
Mail list logo