Details

    • Type: Bug Bug
    • Status: Resolved
    • Priority: Major Major
    • Resolution: Duplicate
    • Fix Version/s: 1.1.3
    • Component/s: None
    • Labels:
      None
    • Environment:

      Ubuntu LTS 12.04, Java Sun Oracle 6 (jre1.6.0_33)

      Description

      It seems that after a while, our keyspaces can't be modified. They accept data, read and write works, truncate works, just keyspace and every column family inside it can't be dropped, created or altered.

      This happens both on our 3 nodes test cluster and on our single-node dev cluster.

      With 1.1.1, a rolling restart of the cluster seemed to solve the issue. This no longer happens with 1.1.2 .

      user@server:~$ cqlsh -3 -k goh_master cassandra1
      Connected to GOH Cluster at cassandra1:9160.
      [cqlsh 2.2.0 | Cassandra 1.1.2 | CQL spec 3.0.0 | Thrift protocol 19.32.0]
      Use HELP for help.
      cqlsh:goh_master> drop columnfamily agents_blueprints;
      cqlsh:goh_master>
      user@server:~$ cqlsh -3 -k goh_master cassandra1
      Connected to GOH Cluster at cassandra1:9160.
      [cqlsh 2.2.0 | Cassandra 1.1.2 | CQL spec 3.0.0 | Thrift protocol 19.32.0]
      Use HELP for help.
      cqlsh:goh_master> DESCRIBE COLUMNFAMILY agents_blueprints

      CREATE TABLE agents_blueprints (
      agent_id ascii,
      archetype ascii,
      proto_id ascii,
      PRIMARY KEY (agent_id, archetype)
      ) WITH COMPACT STORAGE AND
      comment='' AND
      caching='KEYS_ONLY' AND
      read_repair_chance=0.100000 AND
      gc_grace_seconds=864000 AND
      min_compaction_threshold=4 AND
      max_compaction_threshold=32 AND
      replicate_on_write='true' AND
      compaction_strategy_class='SizeTieredCompactionStrategy' AND
      compression_parameters:sstable_compression='SnappyCompressor';

      cqlsh:goh_master>

      1. system.log
        1.33 MB
        Marco Matarazzo

        Activity

        Hide
        Marco Matarazzo added a comment -

        Pavel, maybe you want to look at CASSANDRA-4472.

        Show
        Marco Matarazzo added a comment - Pavel, maybe you want to look at CASSANDRA-4472 .
        Hide
        Pavel Yaskevich added a comment -

        Resolving as duplicate to CASSANDRA-4432 until further notice.

        Show
        Pavel Yaskevich added a comment - Resolving as duplicate to CASSANDRA-4432 until further notice.
        Hide
        Marco Matarazzo added a comment -

        Absolutely OK for me.

        Show
        Marco Matarazzo added a comment - Absolutely OK for me.
        Hide
        Pavel Yaskevich added a comment -

        How about we resolve this one for now and you re-open when it would happen again?

        Show
        Pavel Yaskevich added a comment - How about we resolve this one for now and you re-open when it would happen again?
        Hide
        Marco Matarazzo added a comment -

        It seems that updating the schema from a second node (and maybe the millions of nodetool repair/cleanup/everything) "unlocked" everything.

        As soon as the problem happens again (and I expect it soon), I will more than happily try to apply the patch and see if it solves the problem.

        Show
        Marco Matarazzo added a comment - It seems that updating the schema from a second node (and maybe the millions of nodetool repair/cleanup/everything) "unlocked" everything. As soon as the problem happens again (and I expect it soon), I will more than happily try to apply the patch and see if it solves the problem.
        Hide
        Pavel Yaskevich added a comment -

        Marco, can you try to apply CASSANDRA-4432 and test? I bet that is the problem...

        Show
        Pavel Yaskevich added a comment - Marco, can you try to apply CASSANDRA-4432 and test? I bet that is the problem...
        Hide
        Marco Matarazzo added a comment -

        Update: on the 3-nodes cluster, executing the same DROP on a second node, does the magic. It works with CREATE and ALTER too.

        It can be useful to know that the keyspace has a replication factor of 3.

        Show
        Marco Matarazzo added a comment - Update: on the 3-nodes cluster, executing the same DROP on a second node, does the magic. It works with CREATE and ALTER too. It can be useful to know that the keyspace has a replication factor of 3.

          People

          • Assignee:
            Pavel Yaskevich
            Reporter:
            Marco Matarazzo
          • Votes:
            1 Vote for this issue
            Watchers:
            4 Start watching this issue

            Dates

            • Created:
              Updated:
              Resolved:

              Development