It would help if you provide a bit more detail. For example, you mention an OS upgrade - but what OS? from what version to what version? Do you see any errors in the broker logs at the time at which they shut down?
-- Robin Moffatt | Developer Advocate | ro...@confluent.io | @rmoff On Mon, 14 Jan 2019 at 04:24, Kaushik Nambiar <kaushiknambia...@gmail.com> wrote: > Hello, > I am having a 3 node -3 zookeeper cluster. > My Kafka version is 0.11.x.x. > I am seeing that my Kafka servers start and stop frequently. > We can see atleast 25 start-stop activity per minute. > > This abnormal behaviour started to happen after we had an os upgrade for > the underlying VMs. > > The is upgrade was performed non multiple similar clusters.But only one of > the clusters seems to behave in this manner > > Due to the continuous starting and stopping of the machines,I can also see > increased offline partition count and unclean leader elections. > > Any specific reason for this issue ? > Also can you guys suggest a solution ? > > We tried stopping all the Kafka nodes and starting it again. > It would stop okay,but once they start up,we could see the same problem . > > Regards, > Kaushik Nambiar >
