I noticed this when upgrading a cluster from 1.0.1 to 1.0.5.
We moved one node up to 1.0.5 to test everything was ok and a repair ran from
that node. It streamed files with the new minor file version number to nodes
that were 1.0.1
On restarting one of the nodes while still at 1.0.1 we got this…
INFO [SSTableBatchOpen:1] 2011-12-07 21:56:19,885 SSTableReader.java (line
133) Opening /home/cassandra/data/spi/hits_sort-h-15 (112704 bytes)
ERROR [main] 2011-12-07 21:56:19,900 AbstractCassandraDaemon.java (line 354)
Exception encountered during startup
java.lang.RuntimeException: Can't open sstables from the future! Current
version h, found file: /home/cassandra/data/spi/email_mailers-hb-101
at
org.apache.cassandra.db.ColumnFamilyStore.createColumnFamilyStore(ColumnFamilyStore.java:325)
at
org.apache.cassandra.db.ColumnFamilyStore.createColumnFamilyStore(ColumnFamilyStore.java:303)
at org.apache.cassandra.db.Table.initCf(Table.java:372)
at org.apache.cassandra.db.Table.<init>(Table.java:320)
at org.apache.cassandra.db.Table.open(Table.java:121)
at org.apache.cassandra.db.Table.open(Table.java:104)
at
org.apache.cassandra.service.AbstractCassandraDaemon.setup(AbstractCassandraDaemon.java:183)
at
org.apache.cassandra.service.AbstractCassandraDaemon.activate(AbstractCassandraDaemon.java:337)
at
org.apache.cassandra.thrift.CassandraDaemon.main(CassandraDaemon.java:107)
I think because the 1.0.1 node did not have this see
https://issues.apache.org/jira/browse/CASSANDRA-3478
Perhaps something should be added to NEWS.txt to say all nodes should be on
1.0.3 before running a repair?
Cheers
Aaron
p.s. Hello everyone :)
-----------------
Aaron Morton
Freelance Developer
@aaronmorton
http://www.thelastpickle.com
-----------------
Aaron Morton
Freelance Developer
@aaronmorton
http://www.thelastpickle.com
