I am seeing very strange things when trying to decommission a node in my
cluster (detailed logs attached). Here is a nodetool ring report *after*
decommissioning of node 192.168.4.19 (as seen by any other, properly
functioning node).
192.168.4.15Up Normal 49.9 GB 25.00%
This is rc2 I am assuming?
One thing about remove, the removetoken force command is meant to be run on
the node that originally started a remove and doesn't take a token
parameter. Not relevant to you problem though.
Is this a test cluster and have you tried to reproduce the error? I would be
Just realized the ring output is included in the logs for both of those
nodes. Disregard my earlier request :).
On Wed, Dec 15, 2010 at 8:27 AM, Nick Bailey n...@riptano.com wrote:
This is rc2 I am assuming?
One thing about remove, the removetoken force command is meant to be run on
the
Your logs seem to indicate you have some node flapping problems.
The logs complain about nodes *.18 and *.17 going down and immediately
coming back up, even before the decommission. Is that common throughout
your logs? If thats the case I would attribute seeing *.17 down after the
decommission
email. I need to
decommission two more nodes. I will turn on debug logging when doing so to
try and get some extra info.
Thanks for your help,
Dan
From: Nick Bailey [mailto:n...@riptano.com]
Sent: December-15-10 9:46
To: user@cassandra.apache.org
Subject: Re: Errors when decommissioning