Re: Re-adding a dead node in Cassandra 3.9
We did this but with just two differences it wasn't 3.9 and and always a 3
.X version and we didn't keep the same ip.
We removed the dead node nodetool removenode
Installed Cassandra and configured it to as a new node (we have a playbook
for this) with the same conf that all the others.
Run a repair on each node
Run a cleanup on each node
I found this procedure if you wanna keep the same ip but I never tried it.
On Fri, Oct 26, 2018 at 6:31 AM Ashish Pandey <apandey092@xxxxxxxxx> wrote:
> Our Cassandra cluster uses Cassandra version 3.9. We experienced an issue
> for which documentation doesn't seem clear and I'm hoping someone
> experienced the same issue and we could use some help.
> One of Cassandra node went down and now the host needs to be rebuilt and OS
> needs to be reinstalled and then it can be brought back again to be part of
> the cluster. The host will have same IP as before. What steps should be
> performed for re-adding dead node again? Needless to say, data in the dead
> host after it's back is lost.