Total shot in the dark but could it be related, this talks about CPU but
could have an impact on memory as well:
http://kafka.apache.org/0102/documentation.html#upgrade_10_performance_impact
Hope this helps.
On Sun, 9 Jul 2017 at 10:45 John Yost wrote:
> Hey Ismael,
>
> Thanks a bunch for resp
This is a really common issue with windows. You can try turning off
backups/virus scanning of that folder and sometimes this relieves the
issue. (I assume there is a file attribute you can alter.) It's probably
the case that Kafka itself has the file open, as linux does not have issues
if you mo
I would not mess with internals like this. It's safer if you treat this
like a special case of a Broker expansion.
Don't forget if you are going to have mixed lineage brokers:
*From the documentation*
1. Update server.properties file on all brokers and add the following
properties:
Good Day,
I'm looking for someone to poke holes in my theory.
I want to balance my disk usage across brokers. I want to maintain order
per partition. Yes there are tools but they require manual intervention.
What if created a custom partition strategy. The strategy is to take the
existing part
Look,. I'm a huge fan of sending identical data and using plane old 'wall
time' and averaging a couple runs to make sure you remove any whoops.
You can use fancy tools for reporting but in the real world wall time still
is the most critical factor. And let's face it it's also simple to measure.
You should check the kafka docs. You need to increase your open file
descriptors.
Here's a link that describes how to do
it. (Ignore that this is for hdp the ulimit part is what you need to pay
attend to.)
You can also check kafka docs too they say to change open file descriptors
to 100,000 I
Can you clear up what you mean by "both leader and replica nodes are down
for a given partition"
If all the nodes are down for a partition you can't move the partition as
you don't have any data to copy. (As this is a fundamental part of
re-balancing and leader election. I don't think it will wor
