Reip doesn't work Riak >= 3.0
martinsumner opened this issue · comments
The riak admin reip
command in the riak-admin script depends on two node_package env.sh scripts, that no longer exist:
The reip command also depends on the check_config
command populating the $CONFIG_ARGS.
As this command normally requires on the node being down, it isn't obvious that it can be safely replaced by starting the node and running riak_kv_console:reip/1
from remote_console (and as this would start the ring manager, and potentially gossip this, it might not be safe).
@Bob-The-Marauder - have you guys hit this, and if so do you have a workaround?
This doesn't fix the problem, but offers an alternative reip_manual
. This command is easier to support, as it removes from the code path the requirement for riak_core to be loaded. However, the operator needs to know their Cluster Name and the directory containing the ring files to use it.