incubator-cassandra-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Yan Chunlu <springri...@gmail.com>
Subject would it possible for this kind of data loss?
Date Mon, 22 Aug 2011 01:43:57 GMT
I was aware of the deleted items might be come back alive without proper
node repair.

how about modified items, for example 'A'=>{1,2,3}.  then 'A'=>{4,5}.   if
that possible 'A' change back to {1,2,3}?

I have encountered this mystery problem after go through a mess procedure
with cassandra nodes, (repair,interrupt repair,restart,remove
migration,schema sstable,disable gossip,disablethrift,flush,change RF from
2->3 then 3-> then 2->3, rebalance the cluster using move, change key cache
and row cache to 0 and change back to 200000) something like this.   Is that
possible that I could somehow make the old version data replace the new
version data?

I have 3 nodes and the client read/write consistency level is QUORUM, I have
changed it to ONE for awhile and changed it back.

Mime
View raw message