You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@cassandra.apache.org by "Michael Shuler (JIRA)" <ji...@apache.org> on 2014/05/13 21:22:15 UTC
[jira] [Resolved] (CASSANDRA-7205) Node never know a table has been
DROP or CREATE if its gossip is disabled while executing this query
[ https://issues.apache.org/jira/browse/CASSANDRA-7205?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Michael Shuler resolved CASSANDRA-7205.
---------------------------------------
Resolution: Duplicate
This does reproduce on 2.0.6, but not on current 2.0 HEAD:
{noformat}
mshuler@hana:~$ ccm create -s -n 3 -v 2.0.6 test206
Current cluster is now: test206
mshuler@hana:~$ ccm node1 nodetool disablegossip
mshuler@hana:~$ ccm node2 nodetool status
Datacenter: datacenter1
=======================
Status=Up/Down
|/ State=Normal/Leaving/Joining/Moving
-- Address Load Owns (effective) Host ID Token Rack
DN 127.0.0.1 68.29 KB 66.7% d8acda0c-4726-4cf2-9aef-8200cf563780 -9223372036854775808 rack1
UN 127.0.0.2 35.87 KB 66.7% 9d9f3348-be14-4b93-a9a7-aa07833f71cb -3074457345618258603 rack1
UN 127.0.0.3 35.86 KB 66.7% 7387e665-a5b7-45e2-82a7-ab7cc4d525ad 3074457345618258602 rack1
mshuler@hana:~$ ccm node2 cqlsh
Connected to test206 at 127.0.0.2:9160.
[cqlsh 4.1.1 | Cassandra 2.0.6-SNAPSHOT | CQL spec 3.1.1 | Thrift protocol 19.39.0]
Use HELP for help.
cqlsh> CREATE KEYSPACE foo WITH replication = {'class': 'SimpleStrategy' , 'replication_factor': 2 };
cqlsh> CREATE TABLE foo.bar ( key text PRIMARY KEY , value text );
cqlsh> INSERT INTO foo.bar (key, value) VALUES ( 'blah', 'blargh');
cqlsh> SELECT * from foo.bar ;
key | value
------+--------
blah | blargh
(1 rows)
cqlsh>
mshuler@hana:~$ ccm node1 nodetool enablegossip
mshuler@hana:~$ ccm node1 cqlsh
Connected to test206 at 127.0.0.1:9160.
[cqlsh 4.1.1 | Cassandra 2.0.6-SNAPSHOT | CQL spec 3.1.1 | Thrift protocol 19.39.0]
Use HELP for help.
cqlsh> SELECT * from foo.bar ;
Bad Request: Keyspace foo does not exist
cqlsh>
{noformat}
{noformat}
mshuler@hana:~$ ccm create -s -n 3 -v git:cassandra-2.0 test20head
Fetching Cassandra updates...
Current cluster is now: test20head
mshuler@hana:~$ ccm node1 nodetool disablegossip
mshuler@hana:~$ ccm node2 nodetool status
Datacenter: datacenter1
=======================
Status=Up/Down
|/ State=Normal/Leaving/Joining/Moving
-- Address Load Tokens Owns (effective) Host ID Rack
DN 127.0.0.1 35.94 KB 1 66.7% a709e76b-aa74-4591-b1e7-c43cde8abe3b rack1
UN 127.0.0.2 35.91 KB 1 66.7% 786c8e16-0ec1-4148-93ce-31e8189a5324 rack1
UN 127.0.0.3 35.91 KB 1 66.7% 354a1aa4-3ea6-4fb5-8684-7e2dc5cf6a9e rack1
mshuler@hana:~$ ccm node2 cqlsh
Connected to test20head at 127.0.0.2:9160.
[cqlsh 4.1.1 | Cassandra 2.0.8-SNAPSHOT | CQL spec 3.1.1 | Thrift protocol 19.39.0]
Use HELP for help.
cqlsh> CREATE KEYSPACE foo WITH replication = {'class': 'SimpleStrategy' , 'replication_factor': 2 };
cqlsh> CREATE TABLE foo.bar ( key text PRIMARY KEY , value text );
cqlsh> INSERT INTO foo.bar (key, value) VALUES ( 'blah', 'blargh');
cqlsh> SELECT * from foo.bar ;
key | value
------+--------
blah | blargh
(1 rows)
cqlsh>
mshuler@hana:~$ ccm node1 nodetool enablegossip
mshuler@hana:~$ ccm node1 cqlsh
Connected to test20head at 127.0.0.1:9160.
[cqlsh 4.1.1 | Cassandra 2.0.8-SNAPSHOT | CQL spec 3.1.1 | Thrift protocol 19.39.0]
Use HELP for help.
cqlsh> SELECT * from foo.bar ;
key | value
------+--------
blah | blargh
(1 rows)
cqlsh>
{noformat}
> Node never know a table has been DROP or CREATE if its gossip is disabled while executing this query
> ----------------------------------------------------------------------------------------------------
>
> Key: CASSANDRA-7205
> URL: https://issues.apache.org/jira/browse/CASSANDRA-7205
> Project: Cassandra
> Issue Type: Bug
> Environment: Cassandra 2.0.6
> Reporter: Zhe Yang
> Assignee: Michael Shuler
>
> I'm using Cassandra 2.0.6 and I have 8 nodes. I'm doing some tests by using operations below:
> disable gossip of node A;
> check the status by nodetool in other node, node A is Down now;
> use cqlsh connecting an "Up" node and create a table;
> enable gossip of node A;
> check the status, all nodes are "Up" now.
> Then I find that node A doesn't know this table has been created. Both its own cql shell and nodetool cfstats tell me the table doesn't exist. Even waiting for a few minutes to get the "eventual consistency" final status, node A still doesn't know this table. And I find if each node knows there is a table but I drop it when one node's gossip is disabled, this node will never know the table has been dropped.
> Is this a bug?
--
This message was sent by Atlassian JIRA
(v6.2#6252)