Do you mean that you want to fix sstable table corrupt error and don't mind the testing data ? You may run nodetool scrub <keyspace name> <table name> or nodetool upgradesstable -a <keyspace name> <table name> ( -a is re-write to current version).
Thanks, James On Mon, May 18, 2020 at 12:54 PM Leena Ghatpande <lghatpa...@hotmail.com> wrote: > Running cassandra 3.7 > our TEST cluster has 6 nodes, 3 in each data center > > replication factor 2 for keyspaces. > > we added 1 new node in each data center for testing making it 8 node > cluster. > > We decided to remove the 2 new nodes from cluster, but instead of > decommission, the admin just deleted the data folder by mistake. > > so we ran a nodetool removenode for the 2 nodes. > ran a cleanup and full repair on all remaining nodes. > > But now the whole cluster is corrupt. data returns inconsistent results > and we are getting corrupt sstable errors > > Is there a way to cleanly recover the data? we do not have a old snapshot. >
