[ClusterLabs] How to force remove a cluster node?

Scott Greenlese swgreenl at us.ibm.com
Wed Apr 19 10:21:08 EDT 2017


Tomas,

Yes, I have an IBM internal build we're using for KVM on System Z.    I
tried the --force option and, while it didn't complain,
it didn't work either (as expected, as per bug
https://bugzilla.redhat.com/show_bug.cgi?id=1225423), so
maybe it is a valid option.

[root at zs95kj VD]# date; pcs cluster node remove zs95KLpcs1 --force
Wed Apr 19 10:14:10 EDT 2017
Error: pcsd is not running on zs95KLpcs1
[root at zs95kj VD]#

Hopefully we'll roll in pcs-0.9.143-15.el7_2.1 with our next release of
KVM.

In the mean time, thanks very much for all the valuable feedback.  I'm good
to go for now with the workaround.

Scott

Scott Greenlese ... KVM on System Z - Solutions Test, IBM Poughkeepsie,
N.Y.
  INTERNET:  swgreenl at us.ibm.com




From:	Tomas Jelinek <tojeline at redhat.com>
To:	users at clusterlabs.org
Date:	04/19/2017 03:25 AM
Subject:	Re: [ClusterLabs] How to force remove a cluster node?



Dne 18.4.2017 v 19:52 Scott Greenlese napsal(a):
> My thanks to both Ken Gaillot and Tomas Jelinek for the workaround. The
> procedure(s) worked like a champ.
>
> I just have a few side comments / observations ...
>
> First - Tomas, in the bugzilla you show this error message on your
> cluster remove command, directing you to use the --force option:
>
> [root at rh72-node1:~]# pcs cluster node remove rh72-node3
> Error: pcsd is not running on rh72-node3, use --force to override
>
> When I issue the cluster remove, I do not get and reference to the
> --force option in the error message:
>
> [root at zs93kl ]# pcs cluster node remove zs95KLpcs1
> Error: pcsd is not running on zs95KLpcs1
> [root at zs93kl ]#
>
> The man page doesn't mention --force at my level.

The man page doesn't mention --force for most commands in which --force
can be used. One shouldn't really make any conclusions from that.

> Is this a feature added after pcs-0.9.143-15.el7_2.ibm.2.s390x ?

The feature has been backported to pcs-0.9.143-15.el7_2.1. I cannot
really check if it is present in pcs-0.9.143-15.el7_2.ibm.2.s390x
because I don't have access to that particular build. Based on the name
I would say it was build internally at IBM. However, if the error
message doesn't suggest using --force, than the feature is most likely
not present in that build.

>
> Also, in your workaround procedure, you have me do: 'pcs cluster
> *localnode*remove <name_of_node_to_be_removed> '.
> However, wondering why the 'localnode' option is not in the pcs man page
> for the pcs cluster command?
> The command / option worked great, just curious why it's not
documented ...

It's an internal pcs command which is not meant to be run by users. It
exists mostly for the sake of the current pcs/pcsd architecture ("pcs
cluster node" calls pcsd instance on all nodes over network and pcsd
instance runs "pcs cluster localnode" to do the actual job) and is
likely to be removed in the future. It is useful for the workaround as
the check whether all nodes are running is done in the "pcs cluster
node" command.

Regards,
Tomas

>
> [root at zs93kl # pcs cluster localnode remove zs93kjpcs1
> zs93kjpcs1: successfully removed!
>
> My man page level:
>
> [root at zs93kl VD]# rpm -q --whatprovides /usr/share/man/man8/pcs.8.gz
> pcs-0.9.143-15.el7_2.ibm.2.s390x
> [root at zs93kl VD]#
>
> Thanks again,
>
> Scott G.
>
> Scott Greenlese ... KVM on System Z - Solutions Test, IBM Poughkeepsie,
N.Y.
> INTERNET: swgreenl at us.ibm.com
>
>
> Inactive hide details for Tomas Jelinek ---04/18/2017 09:04:59 AM---Dne
> 17.4.2017 v 17:28 Ken Gaillot napsal(a): > On 04/13/201Tomas Jelinek
> ---04/18/2017 09:04:59 AM---Dne 17.4.2017 v 17:28 Ken Gaillot napsal(a):
>> On 04/13/2017 01:11 PM, Scott Greenlese wrote:
>
> From: Tomas Jelinek <tojeline at redhat.com>
> To: users at clusterlabs.org
> Date: 04/18/2017 09:04 AM
> Subject: Re: [ClusterLabs] How to force remove a cluster node?
>
> ------------------------------------------------------------------------
>
>
>
> Dne 17.4.2017 v 17:28 Ken Gaillot napsal(a):
>> On 04/13/2017 01:11 PM, Scott Greenlese wrote:
>>> Hi,
>>>
>>> I need to remove some nodes from my existing pacemaker cluster which
are
>>> currently unbootable / unreachable.
>>>
>>> Referenced
>>>
>
https://access.redhat.com/documentation/en-US/Red_Hat_Enterprise_Linux/7/html/High_Availability_Add-On_Reference/s1-clusternodemanage-HAAR.html#s2-noderemove-HAAR

>>>
>>> *4.4.4. Removing Cluster Nodes*
>>> The following command shuts down the specified node and removes it from
>>> the cluster configuration file, corosync.conf, on all of the other
nodes
>>> in the cluster. For information on removing all information about the
>>> cluster from the cluster nodes entirely, thereby destroying the cluster
>>> permanently, refer to _Section 4.6, “Removing the Cluster
>>> Configuration”_
>>>
> <
https://access.redhat.com/documentation/en-US/Red_Hat_Enterprise_Linux/7/html/High_Availability_Add-On_Reference/s1-clusterremove-HAAR.html#s2-noderemove-HAAR
>.
>>>
>>> pcs cluster node remove /node/
>>>
>>> I ran the command with the cluster active on 3 of the 5 available
>>> cluster nodes (with quorum). The command fails with:
>>>
>>> [root at zs90KP VD]# date;*pcs cluster node remove zs93kjpcs1*
>>> Thu Apr 13 13:40:59 EDT 2017
>>> *Error: pcsd is not running on zs93kjpcs1*
>>>
>>>
>>> The node was not removed:
>>>
>>> [root at zs90KP VD]# pcs status |less
>>> Cluster name: test_cluster_2
>>> Last updated: Thu Apr 13 14:08:15 2017 Last change: Wed Apr 12 16:40:26
>>> 2017 by root via cibadmin on zs93KLpcs1
>>> Stack: corosync
>>> Current DC: zs90kppcs1 (version 1.1.13-10.el7_2.ibm.1-44eb2dd) -
>>> partition with quorum
>>> 45 nodes and 180 resources configured
>>>
>>> Node zs95KLpcs1: UNCLEAN (offline)
>>> Online: [ zs90kppcs1 zs93KLpcs1 zs95kjpcs1 ]
>>> *OFFLINE: [ zs93kjpcs1 ]*
>>>
>>>
>>> Is there a way to force remove a node that's no longer bootable? If
not,
>>> what's the procedure for removing a rogue cluster node?
>>>
>>> Thank you...
>>>
>>> Scott Greenlese ... KVM on System Z - Solutions Test, IBM
> Poughkeepsie, N.Y.
>>> INTERNET: swgreenl at us.ibm.com
>>
>> Yes, the pcs command is just a convenient shorthand for a series of
>> commands. You want to ensure pacemaker and corosync are stopped on the
>> node to be removed (in the general case, obviously already done in this
>> case), remove the node from corosync.conf and restart corosync on all
>> other nodes, then run "crm_node -R <nodename>" on any one active node.
>
> Hi Scott,
>
> It is possible to remove an offline node from a cluster with upstream
> pcs 0.9.154 or RHEL pcs-0.9.152-5 (available in RHEL7.3) or newer.
>
> If you have an older version, here's a workaround:
> 1. run 'pcs cluster localnode remove <nodename>' on all remaining nodes
> 2. run 'pcs cluster reload corosync' on one node
> 3. run 'crm_node -R <nodename> --force' on one node
> It's basically the same procedure Ken described.
>
> See https://bugzilla.redhat.com/show_bug.cgi?id=1225423 for more details.
>
> Regards,
> Tomas
>
> _______________________________________________
> Users mailing list: Users at clusterlabs.org
> http://lists.clusterlabs.org/mailman/listinfo/users
>
> Project Home: http://www.clusterlabs.org
> Getting started: http://www.clusterlabs.org/doc/Cluster_from_Scratch.pdf
> Bugs: http://bugs.clusterlabs.org
>
>
>
>
>
>
> _______________________________________________
> Users mailing list: Users at clusterlabs.org
> http://lists.clusterlabs.org/mailman/listinfo/users
>
> Project Home: http://www.clusterlabs.org
> Getting started: http://www.clusterlabs.org/doc/Cluster_from_Scratch.pdf
> Bugs: http://bugs.clusterlabs.org
>

_______________________________________________
Users mailing list: Users at clusterlabs.org
http://lists.clusterlabs.org/mailman/listinfo/users

Project Home: http://www.clusterlabs.org
Getting started: http://www.clusterlabs.org/doc/Cluster_from_Scratch.pdf
Bugs: http://bugs.clusterlabs.org



-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.clusterlabs.org/pipermail/users/attachments/20170419/c9638def/attachment-0003.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: graycol.gif
Type: image/gif
Size: 105 bytes
Desc: not available
URL: <https://lists.clusterlabs.org/pipermail/users/attachments/20170419/c9638def/attachment-0003.gif>


More information about the Users mailing list