PDA

View Full Version : Unable to Configure STONITH on a 2-Node Cluster



gaurav
27-Jan-2014, 15:07
I am configuring 2 nodes cluster by installing SLES10 sp4 as OS on both the nodes.
My setup Environment includes:
a). Two Servers(Installed with SLES10 SP4)
b). FC Switch
c). SAN storage
I have performed the configurations as follows:
Step 1: Installed VNC on both the nodes to take the GUI sessions

Step 2: I have installed OES2SP3 and NTP on both the nodes using YAST.

Step 3: I have configured the hacluster pwd on both the nodes by running the following command on both the nodes.
# passwd hacluster

Step 4: I have installed SYSSTAT tool on both the nodes using Yast.

Step 5:I have edit the Kernel file (/etc/sysconfig/kernel) by adding the following line:

INITRD_MODULES="ata_piix megaraid_sas processor thermal fan reiserfs edd lpfc"

Step 5: I have configured Password less SSH on both the nodes.

Step 6: I have performed the heartbeat configuration using Yast on Node-1 and Node-2

Now when I am trying to add STONITH as a cluster resource using Linux HA management (hb_gui)

1. Add items-> native->
2. Resource_id: ssh_stonith
3. Name : ssh
4. Class provide: external/ssh
5. Hostlist: node-1 node-2
6. Using clone

I am able to add the resource, but not able to start it.
Even after clicking on the play button the resource remains offline.

I would like to know whether the above steps which I am following to configure the 2-Node Cluster is proper or not?

Kindly guide me with proper connection diagram and configuration steps.

Thanks in advance!
Gaurav

smflood
27-Jan-2014, 16:34
On 27/01/2014 14:14, gaurav wrote:

> I am configuring 2 nodes cluster by installing SLES10 sp4 as OS on both
> the nodes.
> My setup Environment includes:
> a). Two Servers(Installed with SLES10 SP4)
> b). FC Switch
> c). SAN storage
> I have performed the configurations as follows:
> Step 1: Installed VNC on both the nodes to take the GUI sessions
>
> Step 2: I have installed OES2SP3 and NTP on both the nodes using YAST.
>
> Step 3: I have configured the hacluster pwd on both the nodes by running
> the following command on both the nodes.
> # passwd hacluster
>
> Step 4: I have installed SYSSTAT tool on both the nodes using Yast.
>
> Step 5:I have edit the Kernel file (/etc/sysconfig/kernel) by adding the
> following line:
>
> INITRD_MODULES="ata_piix megaraid_sas processor thermal fan reiserfs edd
> lpfc"
>
> Step 5: I have configured Password less SSH on both the nodes.
>
> Step 6: I have performed the heartbeat configuration using Yast on
> Node-1 and Node-2
>
> Now when I am trying to add STONITH as a cluster resource using Linux HA
> management (hb_gui)
>
> 1. Add items-> native->
> 2. Resource_id: ssh_stonith
> 3. Name : ssh
> 4. Class provide: external/ssh
> 5. Hostlist: node-1 node-2
> 6. Using clone
>
> I am able to add the resource, but not able to start it.
> Even after clicking on the play button the resource remains offline.
>
> I would like to know whether the above steps which I am following to
> configure the 2-Node Cluster is proper or not?
>
> Kindly guide me with proper connection diagram and configuration steps.

Your mention of OES2 SP3 leads me to ask if the aim here is to end up
with a "fault-tolerant" OES server? If so, the way of doing that is with
Novell Cluster Services included within OES. See
http://www.novell.com/documentation/oes2/clus_admin_lx/data/h4hgu4hs.html#h4hgu4hs

You should also note that both SLES10 SP4 and OES2 SP3 are now out of
General Support with SLES11 SP3 and OES11 SP1 (which installs on SLES11
SP2 - OES11 SP2 which installs on SLES11 SP3 is currently in beta) the
latest versions respectively.

HTH.
--
Simon
SUSE Knowledge Partner

------------------------------------------------------------------------
If you find this post helpful and are logged into the web interface,
please show your appreciation and click on the star below. Thanks.
------------------------------------------------------------------------

gaurav
28-Jan-2014, 06:33
Hi Simon,

Thanks for your quick reply.

>>>Your mention of OES2 SP3 leads me to ask if the aim here is to end up
with a "fault-tolerant" OES server?

Yes, my main aim is to configure a "fault-tolerant" 2-node cluster with the following two important properties:



Failover, and
Failback


>>>You should also note that both SLES10 SP4 and OES2 SP3 are now out ofGeneral Support

Yes, I understand that both the SLES10 SP4 and OES2 SP3 are out of General support now. But our requirement is to use only OES2 for cluster setup.
So, I am bound to use the same. Any help in configuring this setup will be highly appriciable.

I will implement your suggestions in my setup and will update you in case of any issue.

Thanks and Regards!
Gaurav

gaurav
31-Jan-2014, 06:38
I am configuring 2 nodes "fault-tolerant" cluster by installing SLES10 sp4 as OS on both the nodes. I am following steps to do the same:

Step 1: Installed VNC on both the nodes to take the GUI sessions
Step 2: I have installed OES2SP3 and configured e-Directory for Node-1.
Step 3:NTP on both the nodes using YAST.
Step 4: I have configured the hacluster pwd on both the nodes by running the following command on both the nodes.
# passwd hacluster
Step 5: I have installed SYSSTAT tool on both the nodes using Yast.
Step 6:I have edit the Kernel file (/etc/sysconfig/kernel) by adding the following line:
INITRD_MODULES="ata_piix megaraid_sas processor thermal fan reiserfs edd lpfc"
Step 7: I have configured Password less SSH on both the nodes.
Step 8: I have performed the heartbeat configuration using Yast on Node-1 and Node-2
Step 9:Now I need to configure Novell Cluster Services.

I would like to know whether the above steps which I am following to configure the 2-Node Cluster is proper or not?
Also kindly let me know whether I can configure the STONITH after configuring the NCS?

Kindly guide me with proper configuration steps.

Thanks in advance!
Gaurav

smflood
31-Jan-2014, 14:53
On 31/01/2014 05:44, gaurav wrote:

> I am configuring 2 nodes "fault-tolerant" cluster by installing SLES10
> sp4 as OS on both the nodes. I am following steps to do the same:
>
> Step 1: Installed VNC on both the nodes to take the GUI sessions
> Step 2: I have installed OES2SP3 and configured e-Directory for Node-1.
> Step 3:NTP on both the nodes using YAST.
> Step 4: I have configured the hacluster pwd on both the nodes by running
> the following command on both the nodes.
> # passwd hacluster
> Step 5: I have installed SYSSTAT tool on both the nodes using Yast.
> Step 6:I have edit the Kernel file (/etc/sysconfig/kernel) by adding the
> following line:
> INITRD_MODULES="ata_piix megaraid_sas processor thermal fan reiserfs edd
> lpfc"
> Step 7: I have configured Password less SSH on both the nodes.
> Step 8: I have performed the heartbeat configuration using Yast on
> Node-1 and Node-2
> Step 9:Now I need to configure Novell Cluster Services.
>
> I would like to know whether the above steps which I am following to
> configure the 2-Node Cluster is proper or not?
> Also kindly let me know whether I can configure the STONITH after
> configuring the NCS?
>
> Kindly guide me with proper configuration steps.

I think that you've misunderstood my original reply - if you want to
create a fault-tolerant OES cluster then you use Novell Cluster Services
and not STONITH (or any other SLES high availability options).

HTH.
--
Simon
SUSE Knowledge Partner

------------------------------------------------------------------------
If you find this post helpful and are logged into the web interface,
please show your appreciation and click on the star below. Thanks.
------------------------------------------------------------------------

gaurav
31-Jan-2014, 15:43
Hi Simon,

Thanks for your quick reply.


I was able to configure the Novell Cluster Services in my setup. But still I am facing the following issues:

Whenever I am trying to run the command "nssmu" to configure the SBD partition, I am getting the "segmentation fault" error message.
Also during OES installation in my "Open Enterprise Server Configuration" window on Node-2, I am getting the following output:

LDAP Server:
NSS Admin Name :cn=node2admin.cn=admin.o=xxxxx

Whereas I have already configured the LDAP server on Node2.

During the OES installation I got the following error on node-2:
In "Saving Linux User Management Configuration" window:
Unable to connect to LDAP server xx.xx.xx.xx or the specified userc=admin,o=xxxx does not have enough privileges toconfigure Linux User ManagementPlease correct the problem and re-run namconfig after the install
unable to LUM add/modify 'admingroup'
unable to LUM add/modify 'admin'

In "Saving Novell Storage Services Configuration" window:
NSS Admin install failed creating eDirectory objects

I am not able to figure out where I have made mistake in installing the OES.

Also do I need to reconfigure the NSS confiuration, if so, please help in the steps of doing it.

Thanks and Regards!
Gaurav

gaurav
07-Feb-2014, 08:24
Hi,

I was able to resolve the above OES installation issue.

What I did is that I re-installed the SLES 10 SP4 on both the nodes and then I reinstalled the OES2 SP3 on both the nodes.

But I am still facing the following issues:



The problem occurs when I install OES. Initially I am hit with the following problems


evms-ha-2.5.5-24.76.4.x86_64 conflicts with other resolvables
-evms-ha-2.5.5-24.76.4.x86_64 conflicts with: novell-cluster-services-1.8.8.723-0.7.x86_64[Novell OES2-SP3]
-Conflict Resolution:
( ) delete evms-ha
delete evms-ha-2.5.5-24.76.4.x86_64
( ) do not install novell-cluster-services
do not install novell-cluster-services-1.8.8.723-0.7.x86_64[Novell OES2-SP3]
( ) Ignore this conflict of evms-ha
pattern:ha_sles-10-51.54.16.x86_64 cannot be installed due to missing dependencies
-There are no installable providers of evms-ha for pattern:ha_sles-10-51.54.16.x86_64
-Conflict Resolution:
( ) delete ha_sles
delete pattern:ha_sles-10-51.54.16.x86_64
( ) Ignore this requirement just here
pattern:xen_server-10-51.54.16.x86_64 conflicts with other resolvables
-pattern:xen_server-10-51.54.16.x86_64 conflicts with pattern: novell-nss-10-141.x86_64[Novell OES2-SP3]
-Conflict Resolution:
( ) delete xen_server
delete pattern:xen_server-10-51.54.16.x86_64
( ) do not install novell-nss
do not install pattern: novell-nss-10-141.x86_64[Novell OES2-SP3]

( ) Ignore this rconflict of xen_server



So, I initially chose ignore for all three options and the cluster came up fine. I was also able to join the second node to the cluster and all seemed to be going smoothly. iManager allowed me to manage the cluster, and then I set about trying to create a shared resource.


NSSMU was no longer functioning, and kept giving 'segmentation fault' errors.


2. I find that if I shutdown (unload) NCS I can manage NSS fine with both imanager and nssmu, however as soon as NCS is running I get segmentation errors on NSSMU and CIMON errors on imanager.
But when I create a new Pool on the Storage Array NSSMU claims that CVB has not registered for pool events.

Exact Message:
21710 CVB has not registered for pool events.

I am not able to understand this error und how to solve it. How can I create new Pools for new Volumes?


Thanks and Regards!
Gaurav

David Gersic
07-Feb-2014, 15:37
On Fri, 07 Feb 2014 07:34:01 +0000, gaurav wrote:

> I was able to resolve the above OES installation issue.

Since this is (no longer) about SLES/HA, you'd probably be better off
asking in the OES support forums, where you'll find people more familiar
with OES and NCS clustering.


--
--------------------------------------------------------------------------
David Gersic dgersic_@_niu.edu
Knowledge Partner http://forums.netiq.com

Please post questions in the forums. No support provided via email.
If you find this post helpful, please click on the star below.