You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@cassandra.apache.org by "Jonathan Ellis (JIRA)" <ji...@apache.org> on 2014/07/30 19:22:40 UTC

[jira] [Commented] (CASSANDRA-7648) ALTER TABLE does not take effect until nodes are restarted after upgrade from 1.1 -> 1.2

    [ https://issues.apache.org/jira/browse/CASSANDRA-7648?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14079576#comment-14079576 ] 

Jonathan Ellis commented on CASSANDRA-7648:
-------------------------------------------

Is the upgrade step required?  What happens if you create from cli directly on 1.2?  What about 2.0?

> ALTER TABLE does not take effect until nodes are restarted after upgrade from 1.1 -> 1.2
> ----------------------------------------------------------------------------------------
>
>                 Key: CASSANDRA-7648
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-7648
>             Project: Cassandra
>          Issue Type: Bug
>         Environment: Mac OSX
>            Reporter: Philip Thompson
>            Priority: Trivial
>
> A 3 node cluster is started on C* 1.1.12. The following commands are issued over cli:
> {code}
> CREATE KEYSPACE upgrade with placement_strategy = 
> 'org.apache.cassandra.locator.SimpleStrategy'
> and strategy_options = {replication_factor:2};
> use upgrade;
> CREATE COLUMN FAMILY cf WITH comparator = UTF8Type
> AND key_validation_class=IntegerType
> AND column_metadata = [
> {column_name: v, validation_class: UTF8Type, index_type: KEYS}];
> {code}
> The cluster is then upgraded to 1.2.18. Both of these versions are installed from source. When the cluster is accessed via cqlsh on 1.2, the CQL column names for table cf are 'key', and 'v'.
> According to http://www.datastax.com/dev/blog/thrift-to-cql3, it should be possible to run
> {code}
> ALTER TABLE cf RENAME key to k;
> {code}
> After performing the alter table, running 'DESC TABLE' shows that the column name has been successfully renamed to 'k' from 'key'. However, if
> {code}
> Select k from cf;
> {code}
> is run, then  'Bad Request: Undefined name k in selection clause' is returned. If the user runs 'select * from cf', the column name shows as 'key' not 'k'. The user is still able to issue queries against 'key', even though 'DESC TABLE' shows 
> {code}
> CREATE TABLE cf (
>   k varint PRIMARY KEY,
>   v text
> ) WITH COMPACT STORAGE AND
>   bloom_filter_fp_chance=0.010000 AND
>   caching='KEYS_ONLY' AND
>   comment='' AND
>   dclocal_read_repair_chance=0.000000 AND
>   gc_grace_seconds=864000 AND
>   read_repair_chance=0.100000 AND
>   replicate_on_write='true' AND
>   populate_io_cache_on_flush='false' AND
>   compaction={'class': 'SizeTieredCompactionStrategy'} AND
>   compression={};
> {code}
> If the cluster is restarted, it fixes the problem.



--
This message was sent by Atlassian JIRA
(v6.2#6252)