[
https://issues.apache.org/jira/browse/CASSANDRA-8487?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14572633#comment-14572633
]
Herbert Kruitbosch edited comment on CASSANDRA-8487 at 6/4/15 12:03 PM:
------------------------------------------------------------------------
I reproduced this for a 4-node with cassandra version **2.1.2**:
{code}
[herbert.kruitbosch@xyz ~]$ cqlsh zyx
Connected to Cassandra productie at zyx:9160.
[cqlsh 4.1.1 | Cassandra 2.1.2 | CQL spec 3.1.1 | Thrift protocol 19.39.0]
Use HELP for help.
cqlsh> select * from system.schema_columns where keyspace_name='system';
(0 rows)
{code}
But not on a similar 4-node system with cassandra version **2.0.9**:
{code}
[herbert.kruitbosch@zyx ~]$ cqlsh xyz
Connected to Cassandra development at xyz
[cqlsh 4.1.1 | Cassandra 2.0.9 | CQL spec 3.1.1 | Thrift protocol 19.39.0]
Use HELP for help.
cqlsh> select * from system.schema_columns where keyspace_name='system';
keyspace_name | columnfamily_name | column_name |
component_index | index_name | index_options | index_type | type |
validator
---------------+-------------------------+-----------------------------+-----------------+------------+---------------+------------+----------------+-----------------------------------------------------------------------------------------------------------------------------
system | IndexInfo | |
null | null | null | null | compact_value |
org.apache.cassandra.db.marshal.BytesType
...
(110 rows)
{code}
Moreover, I get these messages whenever I insert timestamps included records
using the python client:
{code}
Building table metadata with no column meta for system.IndexInfo
{code}
was (Author: herbert.kruitbosch):
I reproduced this for a 4-node with cassandra version **2.1.2**:
{code}
[herbert.kruitbosch@xyz ~]$ cqlsh zyx
Connected to Cassandra productie at zyx:9160.
[cqlsh 4.1.1 | Cassandra 2.1.2 | CQL spec 3.1.1 | Thrift protocol 19.39.0]
Use HELP for help.
cqlsh> select * from system.schema_columns where keyspace_name='system';
(0 rows)
{code}
But not on a similar 4-node system with cassandra version **2.0.9**:
{code}
[herbert.kruitbosch@zyx ~]$ cqlsh xyz
Connected to Cassandra development at xyz
[cqlsh 4.1.1 | Cassandra 2.0.9 | CQL spec 3.1.1 | Thrift protocol 19.39.0]
Use HELP for help.
cqlsh> select * from system.schema_columns where keyspace_name='system';
keyspace_name | columnfamily_name | column_name |
component_index | index_name | index_options | index_type | type |
validator
---------------+-------------------------+-----------------------------+-----------------+------------+---------------+------------+----------------+-----------------------------------------------------------------------------------------------------------------------------
system | IndexInfo | |
null | null | null | null | compact_value |
org.apache.cassandra.db.marshal.BytesType
...
(110 rows)
{code}
> system.schema_columns sometimes missing for 'system' keyspace
> -------------------------------------------------------------
>
> Key: CASSANDRA-8487
> URL: https://issues.apache.org/jira/browse/CASSANDRA-8487
> Project: Cassandra
> Issue Type: Bug
> Components: Core
> Reporter: Adam Holmberg
> Assignee: Aleksey Yeschenko
> Priority: Minor
> Fix For: 2.1.x, 2.2.x
>
>
> Occasionally a Cassandra node will have missing schema_columns information
> where keyspace_name='system'.
> {code}
> cqlsh> select * from system.schema_columns where keyspace_name='system';
> keyspace_name | columnfamily_name | column_name
> ---------------+-------------------+-------------
> (0 rows)
> {code}
> All keyspace and column family schema info is present for 'system' -- it's
> only the column information missing.
> This can occur on an existing cluster following node restart. The data
> usually appears again after bouncing the node.
> This is impactful to client drivers that expect column meta for configured
> tables.
> Reproducible in 2.1.2. Have not seen it crop up in 2.0.11.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
