The config check fits better with the joinSeedNodes process. It’s done
before the actual joining is done. I’d recommend to always use
joinSeedNodes rather than join, also for manual joining. It can be a list
with one entry.
We should probably change Akka management to use joinSeedNodes instead of
Thank you Patrick, I see this behavior is expected.
Considering joining a cluster manually, what is the difference between
the joinSeedNodes and the join processes?
What makes the join process so different that nodes' configurations do
not need to be checked for compatibility?
Thanks, Fred
On
It’s only part of the joinSeedNodes process. Perhaps that is missing in
docs, and contribution fixing that is welcome.
/Patrik
lör 24 feb. 2018 kl. 18:13 skrev Frederic :
> Hello all,
>
> I've just done a few tests with the new Configuration Compatibility Check
> and
Hello all,
I've just done a few tests with the new Configuration Compatibility Check
and noticed that no compatibility check is performed (aka checks are
ignored) when joining the cluster manually using
Cluster(system).join(address). Is that expected?
I didn't find any documentation