[
https://issues.apache.org/jira/browse/CASSANDRA-6053?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13851994#comment-13851994
]
Ryan McGuire commented on CASSANDRA-6053:
-----------------------------------------
OK, reproduced this by killing -9 one of the nodes and then doing a 'nodetool
removenode':
{code}
01:20 PM:~$ kill -9 18961 (PID of node1)
01:21 PM:~$ ccm node1 status
Failed to connect to '127.0.0.1:7100': Connection refused
01:21 PM:~$ ccm node2 status
Datacenter: datacenter1
=======================
Status=Up/Down
|/ State=Normal/Leaving/Joining/Moving
-- Address Load Owns Host ID Token
Rack
DN 127.0.0.1 62.93 KB 20.0% 896644af-8640-4be6-a3ff-e8ed559d851c
-9223372036854775808 rack1
UN 127.0.0.2 51.17 KB 20.0% d3801466-d36d-428c-b4e5-05ff69fe36c0
-5534023222112865485 rack1
UN 127.0.0.3 62.78 KB 20.0% cb36c3ad-df45-4f77-bff5-ca93c504ec08
-1844674407370955162 rack1
UN 127.0.0.4 51.17 KB 20.0% 89031a05-a3f6-4ac7-9d29-6caa0c609dbc
1844674407370955161 rack1
UN 127.0.0.5 51.27 KB 20.0% 4909d856-a86e-493a-a7d0-7570d71eb9d8
5534023222112865484 rack1
# Issue removenode on node3 :
01:21 PM:~$ ~/.ccm/t/node1/bin/nodetool -p 7300 removenode
896644af-8640-4be6-a3ff-e8ed559d851c
01:22 PM:~$ ccm node3 cqlsh
Connected to t at 127.0.0.3:9160.
[cqlsh 4.1.0 | Cassandra 2.0.3-SNAPSHOT | CQL spec 3.1.1 | Thrift protocol
19.39.0]
Use HELP for help.
cqlsh> select * from system.peers;
peer | data_center | host_id | preferred_ip
| rack | release_version | rpc_address | schema_version
| tokens
-----------+-------------+--------------------------------------+--------------+-------+-----------------+-------------+--------------------------------------+--------------------------
127.0.0.2 | datacenter1 | d3801466-d36d-428c-b4e5-05ff69fe36c0 | null
| rack1 | 2.0.3-SNAPSHOT | 127.0.0.2 | d133398f-f287-3674-83af-a1b04ee29f1f
| {'-5534023222112865485'}
127.0.0.5 | datacenter1 | 4909d856-a86e-493a-a7d0-7570d71eb9d8 | null
| rack1 | 2.0.3-SNAPSHOT | 127.0.0.5 | d133398f-f287-3674-83af-a1b04ee29f1f
| {'5534023222112865484'}
127.0.0.4 | datacenter1 | 89031a05-a3f6-4ac7-9d29-6caa0c609dbc | null
| rack1 | 2.0.3-SNAPSHOT | 127.0.0.4 | d133398f-f287-3674-83af-a1b04ee29f1f
| {'1844674407370955161'}
(3 rows)
# Check node2 peers table:
01:23 PM:~$ ccm node2 cqlsh
Connected to t at 127.0.0.2:9160.
[cqlsh 4.1.0 | Cassandra 2.0.3-SNAPSHOT | CQL spec 3.1.1 | Thrift protocol
19.39.0]
Use HELP for help.
cqlsh> select * from system.peers;
peer | data_center | host_id | preferred_ip
| rack | release_version | rpc_address | schema_version
| tokens
-----------+-------------+--------------------------------------+--------------+-------+-----------------+-------------+--------------------------------------+--------------------------
127.0.0.3 | datacenter1 | cb36c3ad-df45-4f77-bff5-ca93c504ec08 | null
| rack1 | 2.0.3-SNAPSHOT | 127.0.0.3 | d133398f-f287-3674-83af-a1b04ee29f1f
| {'-1844674407370955162'}
127.0.0.1 | null | 896644af-8640-4be6-a3ff-e8ed559d851c | null
| null | null | 127.0.0.1 | null
| null
127.0.0.5 | datacenter1 | 4909d856-a86e-493a-a7d0-7570d71eb9d8 | null
| rack1 | 2.0.3-SNAPSHOT | 127.0.0.5 | d133398f-f287-3674-83af-a1b04ee29f1f
| {'5534023222112865484'}
127.0.0.4 | datacenter1 | 89031a05-a3f6-4ac7-9d29-6caa0c609dbc | null
| rack1 | 2.0.3-SNAPSHOT | 127.0.0.4 | d133398f-f287-3674-83af-a1b04ee29f1f
| {'1844674407370955161'}
(4 rows)
01:23 PM:~$ ccm node2 status
Datacenter: datacenter1
=======================
Status=Up/Down
|/ State=Normal/Leaving/Joining/Moving
-- Address Load Owns Host ID Token
Rack
UN 127.0.0.2 51.17 KB 40.0% d3801466-d36d-428c-b4e5-05ff69fe36c0
-5534023222112865485 rack1
UN 127.0.0.3 62.78 KB 20.0% cb36c3ad-df45-4f77-bff5-ca93c504ec08
-1844674407370955162 rack1
UN 127.0.0.4 51.17 KB 20.0% 89031a05-a3f6-4ac7-9d29-6caa0c609dbc
1844674407370955161 rack1
UN 127.0.0.5 51.27 KB 20.0% 4909d856-a86e-493a-a7d0-7570d71eb9d8
5534023222112865484 rack1
{code}
By issuing the removenode on node3, node3 seems to know about the node being
removed and it's peers table is correct. node2, although it's status output
shows node1 going away, it's peers table has not been updated.
> system.peers table not updated after decommissioning nodes in C* 2.0
> --------------------------------------------------------------------
>
> Key: CASSANDRA-6053
> URL: https://issues.apache.org/jira/browse/CASSANDRA-6053
> Project: Cassandra
> Issue Type: Bug
> Components: Core
> Environment: Datastax AMI running EC2 m1.xlarge instances
> Reporter: Guyon Moree
> Assignee: Brandon Williams
> Attachments: peers
>
>
> After decommissioning my cluster from 20 to 9 nodes using opscenter, I found
> all but one of the nodes had incorrect system.peers tables.
> This became a problem (afaik) when using the python-driver, since this
> queries the peers table to set up its connection pool. Resulting in very slow
> startup times, because of timeouts.
> The output of nodetool didn't seem to be affected. After removing the
> incorrect entries from the peers tables, the connection issues seem to have
> disappeared for us.
> Would like some feedback on if this was the right way to handle the issue or
> if I'm still left with a broken cluster.
> Attached is the output of nodetool status, which shows the correct 9 nodes.
> Below that the output of the system.peers tables on the individual nodes.
--
This message was sent by Atlassian JIRA
(v6.1.4#6159)
