Hi Gangaiah, Not sure that removal of WAL and cp is a good idea as long as those can keep info about changes that were not applied checkpointed or applied to persistence files yet.
A safe way to copy the data is to turn off the write workloads to the primary cluster, ensure that checkpointing is finished and then copy the data files. Another approach is to replicate changes to another cluster via Kafka-like approaches or use snapshots of GridGain. - Denis On Mon, Dec 2, 2019 at 9:30 AM Gangaiah Gundeboina <[email protected]> wrote: > Hi Guys, > > We have one live cluster(2.6 native persistence) which serving read and > write continuously. Want to setup new cluster using existing cluster data > without stopping. > > After coping the existing cluster data files, removed metastorage, wal, wal > archive and cp data. And then started the all nodes, getting errors while > cluster activation like jvm pause and record search. > > To do this, can some one suggest us what all things need to take care > before > staring nodes in new cluster. > > > Thanks in advance. > > > > -- > Sent from: http://apache-ignite-developers.2346864.n4.nabble.com/ >
