Hello,

There are no guarantees that you can actually copy data storage from the
working cluster under any kind of load. However, if the cluster is idle and
no checkpoint is running at the moment, data can be copied to another folder
and serve as persistent storage for new cluster.

The only thing here that you have missed is the removal of checkpoint
markers on "WAL history is too short" exception:

Ignite has a relation between WAL, WAL archive and CP folders that has to be
maintained: if you remove WAL and WAL archive, you have to remove the third
"cp" folder also, as it contains checkpoint markers that have references to
WAL (which doesn't exist in your case).

The folder located at $IGNITE_HOME/work/db/$NODE_UUID/cp has to be cleaned
up. After this, node should start in normal way.



--
Sent from: http://apache-ignite-users.70518.x6.nabble.com/

Reply via email to