There is "resumable bootstrap" in 3.0 and newer (maybe 2.2 and newer?), but I've never used it and have no opinion about whether or not I'd trust it myself. I'd personally stop the joining instance, clear the data, and start again.
On Wed, May 1, 2019 at 10:44 AM Nick Hatfield <nick.hatfi...@metricly.com> wrote: > Hello, > > > > Scenario: > > I am joining a new host to my cluster and while joining, one of the hosts > that is streaming data to it, fell over from diskspace filling up. Once the > node was marked DN by the rest of the cluster, I gracefully restarted the > node. I can see all appears to be fine when it came back online, but the > data it was streaming never re-started. > > > > > > Question: > > Is there an easy way, other than stopping and restarting the entire join > process over, to get a node to re-initiate data that it is responsible for > streaming after it dies and comes back online? > > > > Thanks, > > >
