Thanks Jason for the KIP. +1 (non-binding). I assume this change will be added to the upgrade notes as the new values are effective for clusters that were using earlier defaults. Cluster may get into the race condition of having lower replica.lag.time.max.ms on a leader than zookeeper.session.timeout on a follower replica till all the brokers are upgraded but that would not break any semantics.
On Tue, Oct 22, 2019 at 2:20 PM M. Manna <manme...@gmail.com> wrote: > > +1 binding. This kip will possibly fix a lot of extra line of configuration. > > Thanks, > > On Tue, 22 Oct 2019 at 09:06, Stanislav Kozlovski <stanis...@confluent.io> > wrote: > > > Solid KIP. This also has the side-effect of bumping > > zookeeper.connection.timeout.ms to a more reasonable default (the 18s > > session timeout). Perhaps worth mentioning > > > > +1 (non-binding) > > > > On Tue, Oct 22, 2019 at 5:47 AM Ismael Juma <isma...@gmail.com> wrote: > > > > > +1 (binding) > > > > > > On Mon, Oct 21, 2019, 5:28 PM Jason Gustafson <ja...@confluent.io> > > wrote: > > > > > > > I'd like to start a vote for KIP-537: > > > > > > > > > > > > > https://cwiki.apache.org/confluence/display/KAFKA/KIP-537%3A+Increase+default+zookeeper+session+timeout > > > > . > > > > > > > > +1 from me > > > > > > > > Thanks, > > > > Jason > > > > > > > > > > > > > -- > > Best, > > Stanislav > >
