Dmitry Kropachev
Dmitry Kropachev
> I'm wondering how to approach handling keyspace replication changes. > > * We can drop tablets for the keyspace after replication is changed. > As far as I understand,...
Just for the record, it was decided not to implement anything to address scenario when replica set is growing at given moment, but rather document it at https://github.com/scylladb/scylladb/pull/23447
@yarongilor , could you please confirm that there are two issues: 1. That driver trying to reconnect to decomissioned node 2. `c-s` get stuck and does not finish in time.
> I did this for 2 commands as an experiment, in additional commit. I like the new version, so unless anyone has different opinion I'll convert the rest of the...
We need to have it tested, at least manually before merging it.
I have tried to do the same when I worked on rust driver ccm integration. I can tell that manual provisioning (i.e. when you create cluster then add node and...
@fee-mendes, original problem is solved in https://github.com/scylladb/scylla-rust-driver/issues/1023, and verbose logging is going to be implemented at https://github.com/scylladb/scylla-rust-driver/issues/462
> @dkropachev what's that ? Whant to see if it is possible to sync with upstream
> > > @dkropachev what's that ? > > > > > > Whant to see if it is possible to sync with upstream > > What's the point? Be...
On that front I also want to point out that both `3.x` and `4.x` ignore `LoadBalancingPolicy` and return all the replicas in the same order they get it from vnodes...