Open martinsumner opened 1 year ago
https://github.com/basho/riak/pull/1122
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.
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?
Sadly riak_test has no tests for this.