Cluster Administration

Hi All,
I need a quick clarification about the cluster administration of A and B nodes.
Here we have two nodes Node 1 SAP services are running and Node 2 DB services are running both have the cluster administration.
if node 1 failed due to the any issue then all the SAP services like (SAP OSCOL etc) will move to the node 2 automatically Is it(can you explain this) .
and also how we move the SAP services manually to the node2 in case of any issues in node 1.
Regards
Shathishkumar

Hi Sathish
I am Sorry man delay
if node 1 failed due to the any issue then all the SAP services like (SAP OSCOL etc) will move to the node 2 automatically Is it(can you explain this) .
As you said that both nodes are used. Node 1 . SAP instance + SAP Group(DNS Name to connect the SAP system) Node 2 DB & SAP Dialog(Application server) instance. If any one of the physical node fails all the Cluster resource ( Windows fail over Cluster, SAP , DB groups are moving to standby node) are moves to one node with minimum one request time out. in this case you can have fail over Cluster active / active mode. In the Cluster environment only the services are getting failover from one to another node.
In this setup both nodes SAP instance are on,
farther information you can refer the SCN link
High Availability - Frequently Asked Questions
Windows Server Failover Clustering - SAP Business Continuity - SAP Library
Kindly refer the screen shot
BR
SS

Similar Messages

Compressor and Clusters - using Cluster Administrator Cores

So I was recently set the task of setting up our small post production studio's double NAT network and Mac Mini Server, FTP and, most importantly, render farm and compressor (farm?). So we have an iMac, 2 Mac Minis, Mac Mini Server and Mac Pro. I managed (with a great deal of trial and error) to get everything set up, I can go into compressor from any machine and submit a task, and the controller (set as the mac mini server) assigns available cores to start transcoding. However, the MMServer itself ONLY submits 1 core of its' own, and that's set to rendering, not compressor. When I look in Qadmin, all of MMServer's cores assigned to Compressor are greyed out stating "Comprssor (unmanaged)" and can't be used. It is set as Cluster Controller AND SERVICES but it doesn't do the SERVICES part.
Any help would be greatly appreciated! Those extra cores/threads are just sitting there doing nothing!
EDIT - Just to clarify all the machines are set to "Services Only" except the MMServer, and also I do need both Rendering as well as Compressor.

Hey I came in to the office today and re-made the whole cluster again and now it's working. If in doubt... turn it off and on again.

Problem starting managed server in cluster (The name of the managed server is same as admin server)

          We have the following setup.
          CLLUSTER1 and CLUSTER2 are the servers to be in the cluster. CLUSTER3 is the cluster
          administrator. We start CLUSTER3 as a self-managed server and then try and start
          CLUSTER1 and CLUSTER3 as a managed server pointing at CLUSTER3. When we start
          1 and 2 we get the following:
          D:\bea\wlserver6.1\config\CLUSTER1>startmanagedweblogic CLUSTER3 http://10.0.1.1
          03:7001
          D:\bea\wlserver6.1>set PATH=.\bin;C:\WINNT\System32;d:\jdk1.3.1\bin
          D:\bea\wlserver6.1>set CLASSPATH=.;.\lib\weblogic_sp.jar;.\lib\weblogic.jar
          D:\bea\wlserver6.1>echo off
          * To start WebLogic Server, use the password *
          * assigned to the system user. The system *
          * username and password must also be used to *
          * access the WebLogic Server console from a web *
          * browser. *
          D:\bea\wlserver6.1>"d:\bea\jdk131\bin\java" -hotspot -ms64m -mx64m -classpath
          ;.\lib\weblogic_sp.jar;.\lib\weblogic.jar" -Dweblogic.Domain=CLUSTER1 -Dbea.home
          ="d:\bea" -Dweblogic.management.password= -Dweblogic.ProductionModeEnabled=true
          -Dweblogic.Name="CLUSTER3" -Dweblogic.management.server="http://10.0.1.103:7001"
          "-Djava.security.policy==d:\bea\wlserver6.1/lib/weblogic.policy" weblogic.Serve
          r
          <20/11/2001 12:05:10> <Info> <Security> <Getting boot password from user.>
          Enter password to boot WebLogic server:
          Starting WebLogic Server ....
          Connecting to http://10.0.1.103:7001...
          The WebLogic Server did not start up properly.
          Exception raised: weblogic.management.configuration.ConfigurationException: The
          name of the managed server is same as admin server. The managed server will not
          be allowed to start up.
          weblogic.management.configuration.ConfigurationException: The name of the manage
          d server is same as admin server. The managed server will not be allowed to star
          t up.
          at weblogic.management.Admin.getBootstrapLocalServer(Admin.java:998)
          at weblogic.management.Admin.initialize(Admin.java:320)
          at weblogic.t3.srvr.T3Srvr.initialize(T3Srvr.java:359)
          at weblogic.t3.srvr.T3Srvr.run(T3Srvr.java:202)
          at weblogic.Server.main(Server.java:35)
          Reason: Fatal initialization exception
          D:\bea\wlserver6.1>goto finish
          D:\bea\wlserver6.1>cd config\CLUSTER1
          D:\bea\wlserver6.1\config\CLUSTER1>ENDLOCAL
          D:\bea\wlserver6.1\config\CLUSTER1>
          Many thanks in advance,
          Alex Burton
          http://www.e-plus.com.au/


Don't start the admin server as managed. Removing
          -Dweblogic.management.server...
          from the command line should help.
          Alex Burton wrote:
          > We have the following setup.
          > CLLUSTER1 and CLUSTER2 are the servers to be in the cluster. CLUSTER3 is the cluster
          > administrator. We start CLUSTER3 as a self-managed server and then try and start
          > CLUSTER1 and CLUSTER3 as a managed server pointing at CLUSTER3. When we start
          > 1 and 2 we get the following:
          >
          > D:\bea\wlserver6.1\config\CLUSTER1>startmanagedweblogic CLUSTER3 http://10.0.1.1
          > 03:7001
          >
          > D:\bea\wlserver6.1>set PATH=.\bin;C:\WINNT\System32;d:\jdk1.3.1\bin
          >
          > D:\bea\wlserver6.1>set CLASSPATH=.;.\lib\weblogic_sp.jar;.\lib\weblogic.jar
          >
          > D:\bea\wlserver6.1>echo off
          >
          > ***************************************************
          > * To start WebLogic Server, use the password *
          > * assigned to the system user. The system *
          > * username and password must also be used to *
          > * access the WebLogic Server console from a web *
          > * browser. *
          > ***************************************************
          >
          > D:\bea\wlserver6.1>"d:\bea\jdk131\bin\java" -hotspot -ms64m -mx64m -classpath
          > ".
          > ;.\lib\weblogic_sp.jar;.\lib\weblogic.jar" -Dweblogic.Domain=CLUSTER1 -Dbea.home
          > ="d:\bea" -Dweblogic.management.password= -Dweblogic.ProductionModeEnabled=true
          > -Dweblogic.Name="CLUSTER3" -Dweblogic.management.server="http://10.0.1.103:7001"
          > "-Djava.security.policy==d:\bea\wlserver6.1/lib/weblogic.policy" weblogic.Serve
          > r
          > <20/11/2001 12:05:10> <Info> <Security> <Getting boot password from user.>
          > Enter password to boot WebLogic server:
          > Starting WebLogic Server ....
          > Connecting to http://10.0.1.103:7001...
          > ***************************************************************************
          > The WebLogic Server did not start up properly.
          > Exception raised: weblogic.management.configuration.ConfigurationException: The
          > name of the managed server is same as admin server. The managed server will not
          > be allowed to start up.
          > weblogic.management.configuration.ConfigurationException: The name of the manage
          > d server is same as admin server. The managed server will not be allowed to star
          > t up.
          > at weblogic.management.Admin.getBootstrapLocalServer(Admin.java:998)
          > at weblogic.management.Admin.initialize(Admin.java:320)
          > at weblogic.t3.srvr.T3Srvr.initialize(T3Srvr.java:359)
          > at weblogic.t3.srvr.T3Srvr.run(T3Srvr.java:202)
          > at weblogic.Server.main(Server.java:35)
          > Reason: Fatal initialization exception
          > ***************************************************************************
          >
          > D:\bea\wlserver6.1>goto finish
          >
          > D:\bea\wlserver6.1>cd config\CLUSTER1
          >
          > D:\bea\wlserver6.1\config\CLUSTER1>ENDLOCAL
          >
          > D:\bea\wlserver6.1\config\CLUSTER1>
          >
          > Many thanks in advance,
          > Alex Burton
          > http://www.e-plus.com.au/

Installation of SAPCCM4X-Agent on Windows Cluster

Hi!
I woulld like to install SAPCCM4X-Agent on Windows Cluster.
Does anyone know the procedure here (e.g. install the agent twice on both nodes, etc.)...
Thank you very much!
regards
Thom

Hello Thom,
Here is a note that will help you out.
SAP Note 536954 - OS data in a cluster environment for OS07
As per the note
The CCMS agent must be located on a shared disk, so that the cluster administrator software can start the service later.
Regards,
Siddhesh
Edited by: Siddhesh Ghag on May 27, 2008 11:02 AM

How should I start, stop or restart SQL services if it is in a cluster?

Hi,
For migration purpose I need to stop sql server services in a cluster. I don't want to failover the cluster I just want to stop the sql server services before the windows system admin shutdowns the windows server and later start again once migration is done.
This is clone migration so I don't have to do anything except stopping and starting the services of sql. I found two places there to take service offline and online, one is just clicking the services and application and in the right hand side directly right
clicking the instance and take service offline, another is expanding the instance and taking offline agent and sql server service one by one.
I don't know which is the way.
Please some one help me.

Hello,
You can use SQL Server Configuration Manager, SQL Server Management Studio, Cluster Administrator.
Do not use the Services applet in Control Panel.
Hope this helps.
Regards,
Alberto Morillo
SQLCoffee.com

Howto take SQL Server 2005 Cluster AND Cluster Group Offline

I need to update firmware on the backend storage used by our Active/Passive SQL Server 2005 Cluster. The first step in the process states that all I/O to the disks should be stopped. Per that, I need to take the the SQL Server cluster offline AND the cluster group; however, I then need to use one of the servers that participates in the cluster to perform the upgrade on the storage array. Using the process (which doesn't account for having to use one of clustered servers to perform updates) outlined at http://forums.microsoft.com/MSDN/ShowPost.aspx?PostID=1550288&SiteID=1 as a basis, below is my best guess regarding what I will need to do:
Shutdown:
1. Backup all of your databases prior to the outage and ensure they are in a safe place (NOT on the disk drives in the cluster)
2. Take the SQL Server group OFFLINE. This will cleanly shutdown the SQL Server and it will remain offline when you power the machine back on.
3. Once SQL Server is offline, power off the node which does NOT have the cluster resource currently running on it
4. Save the node which is running the cluster group for last - make a note of this machine as it will be the first one you power back on. MY ADDITION: Before powering it down, set the cluster service to manual so that when you restart the server to perform the storage array maintenance, the cluster will not be running - and likewise not generating I/O against the disk array.
MY ADDITION: 5. Restart the server from step 4. Takes the cluster offline "cleanly" and brings the server back on line so that the storage array maintenance can be performed.
6. Update the firmware on the storage array. After that is complete, set the cluster service back to "Automatic".
Start up:
1. Restart the server. Per the previous steps, the server to be restarted will be the server that last hosted the cluster group and the one used to update the firmware on the storage array.
2. RDP into the node, open cluster administrator, and verify that the cluster came back online successfully and everything looks clean
3. Power on the second node.
4. Failover the cluster group across each node in the cluster and validate that everything looks OK
5. Bring your SQL Server groups back online
6. Validate that everything is operational
Is there a better way to do this or official documentation for the process?
Thanks.

FYI: I did use the process outlined above to briing the cluster offline, perform updates, and bring it back online. It worked as I had hoped. As a test, I had also done the same on a test [non-SQL] cluster prior to making the updates to our SQL cluster. Is it the best way to do it? Who knows. Did it work? Yes.

Windows 2003 Cluster IP Conflict during failover. Possible Spilt Brain Syndrome ?

Dear all,
I would like to ask if anyone had a case of IP Conflict on the Virtual Cluster IP Address resource ?
It was observed that both Windows cluster nodes presented the Virtual IP address resource at the same time.
Could I know how can such a thing happen, there are may safeguards to prevent such spilt brain symptom.
Thanks a lot.

Dear Tim & Alex
Apologies for the late reply . Forgot about the alert me feature...
Our switches are configured with a feature to lock-out (disable) any port with a duplicate IP address/MAC address.
May I know would there actually be a time for maybe for a split second where both Virtual Cluster IP addresses are presented on the NICs of both nodes ?
This incident happened again but this time the situation was during a restart of the passive node. What happened was that upon restart, the resources on the active node attempted to fail over to the passive node when I restarted the passive node. Upon login,
I then realised that the port on the restarted node was portlocked and the services stayed on the restarted node, even though It was strange for it to failover as it had not reason to ( it was running fine on the other node) . I tried to open the cluster administrator
on the node that was not restarted and it had cluster administrator in a hanged state however the service was indicated as started. The cluster administrator on the restarted node indicated cluster IP has failed while the ownership was all on the restarted
node.
I find it curious that the cluster administrator would be unresponsive on the node that was not restarted as it did a failover smoothly just 30 min ago.
Thanks for your time!

2003 to 2008 R2 file cluster migration - shares fail to migrate due to missing client access point

Hello,
I need to migrate my 2003 cluster to 2008 R2, but am having difficulty doing so. I have a 2003 2 node disk and node majority cluster which has 3 groups (2 individual file shares) and a separate group for the quorum.
I'm trying to migrate my 2003 groups to a 2 node 2008 r2 cluster.
I've ran through the cluster migration wizard as
http://technet.microsoft.com/en-us/library/ff182337(v=ws.10).aspx), the disk resources within my group migrate, but the file shares don't.
In summary:
- I ran the 2008 r2 cluster run migration wizard and connected to my 2003 cluster
- Once the wizard completes, I left the the newly migrated group on the 2008 r2 cluster offline
- Took the appropriate file share and disk resource group offline on 2003
- Ensured the iSCSI disks for my migrated data are showing correctly on the 2008 R2 node (I'm using the same drive letter which my 2003 node used for the moved disk)
- I then brough the resources online. Resources fail to come online as client access point is missing, so I add a new access point (http://blogs.technet.com/b/askcore/archive/2010/08/24/using-multiple-client-access-points-cap-in-a-windows-server-2008-r2-failover-cluster.aspx)
- Restart my 2008 r2 cluster node
- File share is missing! The file share has not been imported
I then tried the registry key export and import:
- On my 2003 active cluster node I export HKLM\SYSTEM\CurrentControlSet\Services\LanmanServer\Shares
- I take the appropriate cluster resource and disk offline on my 2003 node
- Attach the disk to my 2008 R2 cluster node
- Added the new disk as a cluster resource (again keeping the same drive letter)
- Merged the 2003 registry key with the share information (note examining regedit shows the newly imported key)
- Restart the server, but the share doesn't show up in the cluster administrator or share management tool on my 2008 R2
1. Is it even possible to transfer file shares using the cluster migration wizard from 2003 to 2008 r2? Maybe I'm missing a step with the client access point (or do I need to transfer all cluster resources over at once).
2. Should the registry key export import work, when the shares have been created under the cluster administration console (as opposed to windows explorer)?
Thanks

Hi,
Currently there is no such a tool which could help transfer Sharing permissions to a new server or a new cluster.
Here is an article about file share in a cluster:
Working with File Shares in Windows Server 2008 (R2) Failover Clusters
http://blogs.technet.com/b/askcore/archive/2010/08/19/working-with-file-shares-in-windows-server-2008-r2-failover-clusters.aspx
From this article, share permission of a Windows Server 2008 R2 failover Cluster will be stored in HKLM\Cluster\Resources\<GUID>\Parameters.
You can test to export that keys to new cluster to see if shares will occur however as it is GUID involved, it may not work on a new cluster.
If you have any feedback on our support, please send to [email protected]

J2EE 6.20 Java Cluster - Synchronisation of EP AND J2EE Components

Hello,
<b>
J2EE 6.20 Java Cluster -> Synchronisation of EP AND J2EE Components </b>
We have a productive SAP Enterprise Portal J2EE Cluster Farm with 3 Server. The version of EP is: 6.0 SP2 Patch 28. One State Controller, one Backup Controller and one new Dialog Instance.
We would like to know, which folders, shares and other j2ee/ep-components are synchronized from the state Controller-Server to the new Dialog Instance?
Best Regards,
Ralf Steffens

Thank you for your response.
In the cluster administration, the Server IP column reflects the IP of the OS instance on which the server runs. The question is relative to which LDAP server each server process grabs from a round-robin type resolution.
Any other thoughts?
MMcHugh

Cluster unexpected error

Hello All,
I have installed single node cluster 3.2 in solaris 10. I am getting below error when i am running scstat
*# /usr/cluster/bin/scstat | more*
*scstat: unexpected error.*
*Jan 19 04:21:06 node1 Cluster.CCR: Unable to open door descriptor /var/run/rgmd_receptionist_door*
*-- Cluster Nodes --*
*Node name Status*
*Cluster node: node1 Online*
Could you please explain me what is error about.
Regards,
R. Rajesh Kannan.

Hi ,
The below are output of svcs -x output
bash-3.00# /usr/cluster/bin/scstat -g
Jul 23 10:13:20 Cluster.CCR: Unable to open door descriptor /var/run/rgmd_receptionist_door
scstat: unexpected error.
bash-3.00# hostname
caotslab202
bash-3.00# ps -ef | grep rgmd
root 730 668 0 10:13:28 console 0:00 grep rgmd
bash-3.00# svcs -x
svc:/system/filesystem/local:default (local file system mounts)
State: maintenance since Wed Jul 23 09:53:46 2008
Reason: Start method exited with $SMF_EXIT_ERR_FATAL.
See: http://sun.com/msg/SMF-8000-KS
See: /var/svc/log/system-filesystem-local:default.log
Impact: 56 dependent services are not running. (Use -v for list.)
svc:/network/rpc/gss:default (Generic Security Service)
State: uninitialized since Wed Jul 23 09:24:09 2008
Reason: Restarter svc:/network/inetd:default is not running.
See: http://sun.com/msg/SMF-8000-5H
See: gssd(1M)
Impact: 40 dependent services are not running. (Use -v for list.)
svc:/network/rpc/meta:default (SVM remote metaset services)
State: uninitialized since Wed Jul 23 09:24:07 2008
Reason: Restarter svc:/network/inetd:default is not running.
See: http://sun.com/msg/SMF-8000-5H
See: rpc.metad(1M)
Impact: 31 dependent services are not running. (Use -v for list.)
svc:/application/print/server:default (LP print server)
State: disabled since Wed Jul 23 09:24:09 2008
Reason: Disabled by an administrator.
See: http://sun.com/msg/SMF-8000-05
See: lpsched(1M)
Impact: 1 dependent service is not running. (Use -v for list.)
svc:/network/rpc/scadmd:default (Sun Cluster administrative service)
State: uninitialized since Wed Jul 23 09:26:37 2008
Reason: Restarter svc:/network/inetd:default is not running.
See: http://sun.com/msg/SMF-8000-5H
Impact: 1 dependent service is not running. (Use -v for list.)
svc:/network/rpc/smserver:default (removable media management)
State: uninitialized since Wed Jul 23 09:24:16 2008
Reason: Restarter svc:/network/inetd:default is not running.
See: http://sun.com/msg/SMF-8000-5H
See: rpc.smserverd(1M)
Impact: 1 dependent service is not running. (Use -v for list.)
svc:/network/rpc/rstat:default (kernel statistics server)
State: uninitialized since Wed Jul 23 09:24:13 2008
Reason: Restarter svc:/network/inetd:default is not running.
See: http://sun.com/msg/SMF-8000-5H
See: rpc.rstatd(1M)
See: rstatd(1M)
Impact: 1 dependent service is not running. (Use -v for list.)
bash-3.00# svcs -x
Regards,
R. Rajesh Kannan.

2 node failover cluster power down

I have a 2node failover cluster. When I power down a node that has the SQL server instance and resources, all the resources and service failover to the other node. When I see that all the resources and service report "online" I then power
that node. I am being told that this is improper because failover may not have completed. Is that correct?
Also, in our 2 node failover cluster is there a proper sequence to restarting the powered down nodes?

Hi,
The cluster group containing SQL Server can be configured for automatic failback to the primary node when it becomes available again. By default, this is set to off.
To Configure:
Right-click the group containing SQL Server in the cluster administrator, select 'properties' then 'failback' tab.
To prevent an auto-failback, select 'Prevent Failback', to allow select 'Allow Failback' then one of the following options:
Immediately: Not recommended as it can disrupt clients
Failback between n and n1 hours: allows a controlled failback to a preferred node (if it's online) during a certain period.
The related article:
Windows Failover Clustering Overview
http://blogs.technet.com/b/rob/archive/2008/05/07/failover-clustering.aspx
Hope this helps.
We
are trying to better understand customer views on social support experience, so your participation in this
interview project would be greatly appreciated if you have time.
Thanks for helping make community forums a great place.

Where can I find the cluster home in 9i RAC like CRS_HOME in 10g

I am trying to find out the cluster home for executing command srvctl,crs_stat in oracle 9i RAC . I dont find find cluster home folder on sun solaris.
Thanks

what is the utility for veritas cluster administration in 9i ?
Where is cluster log file stored if dba want to debug ?Check this out for information on VCS - http://www.darklab.net/resources/veritas-cluster-server-debug.html - a very good and usefull resource.
Do we need VIP for 9i RAC with Veritas Cluster ?Oracle doesn't need it, but VCS may.
How is failover to another node work compare to load balancing? Try the Oracle Documentation at http://download.oracle.com/docs/cd/B19306_01/network.102/b14212/toc.htm and just search for Load Balancing and TAF, if has lot of information on those concepts, with examples.
HTH
Thanks
Chandra Pabba

Documentation about SAP on Windows Cluster

Hello!
We are searching for appropriate documentation about SAP on Windows Cluster:
- administration of SAP on Windows Cluster with Oracle
- the process of kernel change
Any helpful information will be very appreciated.
Thank you!

Hello Tom,
I think the following link can be a good starting point for you:
[SAP on Microsoft Windows|SAP on Microsoft Windows]
You will find the info at the bottom of the screen under:
- Important SAP Notes on Windows
- SAP on Windows - Document Archive
Success.
Wim Van den Wyngaert

Repair broken Oracle cluster on Win 2000

I have an Oracle cluster (8.1.7. with OFS) running in a Windows 2000 AS environment. There are two nodes and -luckey me- it is on a test network.
A while ago I tried to install OEM in order to be able to configure a OMS.
I performed the install on the inactive node without stopping the Oracle (cluster)services (stupid, stupid....) and broke it off when it hung at the end.
From that moment on this node is no longer available for failover in the cluster. This operation seems to have affected this node's Oracle home. Later on I performed a de-install but this did'nt help.
Now I wonder:
is it possible to repair this 'broken' cluster and if so what would be the procedure to follow?
Thanks in advance,
Georges

At last I solved the problem myself.
On both nodes I desinstalled all Oracle products (Server, Fail Safe..) and later installed them again (first Oracle EE on both nodes, then OFS on both nodes).
Mind that all the fysical files for my databases were still available on SAN disks!
With the ORADIM utility I created services for my two databases on one node.
I recreated the Oracle clustergroup and added the necessary disk resources etc with Cluster administrator (Win2000).
I started Fail Safe Manager and added (one by one) my databases as standalone resources to the Oracle group and went through the verification process without warning.
Now everything is working fine again...

Kernel Upgrade in MSCS cluster

Hi all,
I have to do Kernel Upgrade in MSCS environment.
Please suggest me how to do?
Regards,
shanker C

Hi Shanker,
Given below is the procedure to perform kernel upgrade in MSCS environment
Activities to be done at OS level
1. Stop all instances and take all the resource offline(including windows cluster service) in cluster administrator.
2. Unpack the archives to a working directory and take backup of usr\sap\<SAPSID>\SYS\exe\uc(nuc)\<Platform>:and replace all old files with the new one.
3. On all MSCS nodes, replace the SAP cluster dlls with those in NTCLUST.SAR
a) Copy ntcluster\saprc.dll %windir%\system32
b) Copy ntcluster\saprcex.dll %windir%\system32
4. On all MSCS nodes, replace the following dlls and exe in the
windir%\SapCluster folder with those in folder \usr\sap\<SAPSID>\SYS\exe\uc(nuc)\<Platform>:
sapevents.dll
sapntchk.exe
sapstartsrv.exe
sapxpg.exe
5. Bring cluster resources online and start the instances.
Hope this answer your question
Regards
Waseem

Cluster Administration

Similar Messages

Maybe you are looking for