cassandra - start accepting writes after nodetool drain

3
votes

Once we do 'nodetool drain', what can we do to make it start accepting writes again?

on node A I did nodetool drain.

On another node B, when I run nodetool repair, it says:

Cannot proceed on repair because a neighbor (/<ip-of-A>) is dead: session failed

any ideas what could be happening here?

thanks.

3
cassandradatastax

3 Answers

1
votes

Nodetool drain is intended to be run before stopping a node. If you don't want to stop the node then you probably want to run some combination of the following commands (depending on what your goal is):

  • nodetool flush
  • nodetool disablethrift
  • nodetool disablegossip
  • nodetool disablebinary
0
votes

For restore the operation, it is enough to restart the cassandra service.

sudo systemctl restart cassandra.service

-1
votes

After executing nodetool drain, once the the status is DOWN & NORMAL, execute <nodetool removenodes UUID> then based on replication factor proceed or nodetool repair. By Replication Factor (RF), I mean the value you would have set for sync of your data on number of nodes.