cassandra-commits mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Jeff Jirsa (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (CASSANDRA-14414) Errors in Supercolumn support in 2.0 upgrade
Date Wed, 25 Apr 2018 03:22:00 GMT

    [ https://issues.apache.org/jira/browse/CASSANDRA-14414?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16451564#comment-16451564
] 

Jeff Jirsa commented on CASSANDRA-14414:
----------------------------------------

2.0 is EOL, and we weren't expecting to cut a new release. What happens if you finish the
rolling upgrade to 2.0? Does it finish (failing reads while it's mixed)? 




> Errors in Supercolumn support in 2.0 upgrade
> --------------------------------------------
>
>                 Key: CASSANDRA-14414
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-14414
>             Project: Cassandra
>          Issue Type: Bug
>            Reporter: Ken Hancock
>            Priority: Major
>
> In upgrading from 1.2.18 to 2.0.17, the following exceptions started showing in cassandra
log files when the 2.0.17 node is chosen as the coordinator.  CL=ALL reads will fail as a
result.
> The following ccm script will create a 3-node cassandra cluster and upgrade the 3rd node
to cassandra 2.0.17
> {code}
> ccm create -n3 -v1.2.17 test
> ccm start
> ccm node1 cli -v -x "create keyspace test with placement_strategy='org.apache.cassandra.locator.SimpleStrategy'
and strategy_options={replication_factor:3}"
> ccm node1 cli -v -x "use test;
> 	create column family super with column_type = 'Super' and key_validation_class='IntegerType'
and comparator = 'IntegerType' and subcomparator = 'IntegerType' and default_validation_class
= 'AsciiType'"
> ccm node1 cli -v -x "use test;
> 	create column family shadow with column_type = 'Super' and key_validation_class='IntegerType'
and comparator = 'IntegerType' and subcomparator = 'IntegerType' and default_validation_class
= 'AsciiType'"
> ccm node1 cli -v -x "use test;
> 	set super[1][1][1]='1-1-1';
> 	set super[1][1][2]='1-1-2';
> 	set super[1][2][1]='1-2-1';
> 	set super[1][2][2]='1-2-2';
> 	set super[2][1][1]='2-1-1';
> 	set super[2][1][2]='2-1-2';
> 	set super[2][2][1]='2-2-1';
> 	set super[2][2][2]='2-2-2';
> 	set super[3][1][1]='3-1-1';
> 	set super[3][1][2]='3-1-2';
> 	"
> ccm flush
> ccm node3 stop
> ccm node3 setdir -v2.0.17
> ccm node3 start
> ccm node3 nodetool upgradesstables
> {code}
> The following python uses pycassa to exercise the range_slice Thrift API:
> {code}
> import pycassa
> from pycassa.pool import ConnectionPool
> from pycassa.columnfamily import ColumnFamily
> from pycassa import ConsistencyLevel
> pool = ConnectionPool('test', server_list=['127.0.0.3:9160'], max_retries=0)
> super = ColumnFamily(pool, 'super')
> print "fails with ClassCastException"
> super.get(1, columns=[1,2], read_consistency_level=ConsistencyLevel.ONE)
> print "fails with RuntimeException: Cannot convert filter to old super column format...""
> super.get(1, column_start=2, column_finish=3, read_consistency_level=ConsistencyLevel.ONE)
> {code}



--
This message was sent by Atlassian JIRA
(v7.6.3#76005)

---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@cassandra.apache.org
For additional commands, e-mail: commits-help@cassandra.apache.org


Mime
View raw message