Mscs cluster failover

Hello,
We have installed production on MSCS cluster.Installation is finished successfully.Cluster resources for ASCS and SCS instance in sapgroup
are not getting failover.If we take SCS instance resource offline and try to failover ascs instance resource manually,it doesnot work out first time.But 2nd time it worked out.Also,with same status if we restart one node,ASCS instance resource doesnt come online.
We tried out many things without success.Plz help ASAP.
Regards,
shital

Hi,
MSCS needs a lot of steps to configure correctly.
First try out with one or two physical disk resources if they are completely failover. then try SAP resources.
I have two nodes (Active and Passive), Active acts as Database instance (mssql), and passive acts as dialog instance (R/3). When passive has problems and shuts off, then the active node takes over the passive resource (R/3) and acts as a central instance. After the passive comes back online, then you have to move the R/3 resource from the active node to the passive manually, not automatically.
Please let me know if you still have problems.
Regards.

Similar Messages

OFFLINE backup and DB13 jobs configuratins error in MSCS cluster with OFS

Hello,
We have recently installed SAP ERP 6.0 EhP4 SR1 in MSCS cluster using Oracle Fail Safe.
The installation is successful and all the failover scenarios work fine.
I need to schedule offline backups and DB13 jobs, which is not possible in current scenario as once the database shuts down it fails over to node2.
I am following sapnote 657999 to install standalone gateway on both the nodes of the cluster and assigned the service to ORACLE<SID> group. the gw<SID> is different from the ERP<SID>.
However, the service SAP<gwSID>_<gwSN> fail to come up with the following error.
The service SAPSGA_30 cannot be started on local computer
error 1067: the process terminated unexpectedly.
The ../SapCluster/sapgw and associated directories have full permission.
Following are the contents of the profiles in ../SapCluster/sapgw
default.pfl
SAPDBHOST = ORACLEPRD
gw30.pfl
SAPSYSTEMNAME = SGA
INSTANCE_NAME = GW30
SAPSYSTEM = 30
DIR_PROFILE=C:\WINDOWS\SapCluster\sapgw
DIR_EXECUTABLE=C:\WINDOWS\SapCluster\sapgw
DIR_INSTANCE=C:\WINDOWS\SapCluster\sapgw
SAPLOCALHOST = ORACLEPRD
SAPLOCALHOSTFULL = ORACLEPRD
startgw.pfl
SAPSYSTEMNAME = SGA
INSTANCE_NAME = GW30
SAPSYSTEM = 30
DIR_EXECUTABLE = C:\WINDOWS\SapCluster\sapgw
DIR_PROFILE = C:\WINDOWS\SapCluster\sapgw
DIR_INSTANCE = C:\WINDOWS\SapCluster\sapgw
_GW=gwrd.exe
Start_Program_00 = local $(DIR_EXECUTABLE)\$(_GW) pf=$(DIR_PROFILE)\gw30.pfl
Now as the service is a part of my ORACLE<ERP SID> group, everytime I try to bring the group online, it fails due to the gw service and toggles many times in the two nodes before coming to rest on one node withfailed gw service and database in the up condition.
Any insight in this regard will be really helpful.
Thanks
Nischal

Hello,
Yes we were able to configure offline backups.
We opted to install standalone gateway on both nodes and added it to the DB group. That provides the essential shell console.
You need to follow the SAP Note 657999 and then 378648 for setting the necessary environment variables.
One piece of advice, use the SID and system no. for the independent gateway, different from any of your SAP installations.
Thanks
Nischal

MSCS Cluster Installation

Hello Experts,
Can anybody help me configure the SAPGUI on the end user's system in a MSCS Cluster installation ?
I have Central Instance on Node 1 and Dialog Instance on Node 2 with SQL Server 2005 on Windows 2003 Enterprise Edition. The SQL and SAP cluster is working absolutely fine, but when there is a failover to Node 2, the user's connection to the server is getting disconnected
Please advice.
Edited by: The Fuhrer on Dec 18, 2008 5:45 AM

Hi kiran
use db2mscs.eee is the template and copy this to db2mscs.cfg.
and edit the file using page number 255.
Global section
DB2_INSTANCE = DB2<SID>
DB2_LOGON_USERNAME = domainname\db2<sid>
DB2_LOGON_PASSWORD = password
CLUSTER_NAME = clustername
MSCS Group for Node 0
GROUP_NAME = DB2 <DBSID> Group
DB2_NODE = 0
IP_NAME = DB2 IP <SID>
IP_ADDRESS=ipaddess of virtual db2 server
IP_SUBNET=subnet mask ipaddress
IP_NETWORK=Public LAN
NETNAME_NAME = DB2 NetName <SID>
NETNAME_VALUE = hostname of virtual db2 server
NETNAME_DEPENDENCY = DB2 IP <SID>
DISK_NAME = Disk L: these 3 are the shared disks
DISK_NAME = Disk M:
DISK_NAME = Disk N:
INSTPROF_DISK = Disk L:
#TARGET_DRVMAP_DISK = Disk M:
MSCS Group for Node 1
#GROUP_NAME = DB2 Node 1
#DB2_NODE = 1
#IP_NAME = IP Address for Node 1
IP_ADDRESS=9.26.118.252
IP_SUBNET=255.255.255.0
IP_NETWORK=Ethernet(2)
#NETNAME_NAME = Network Name for Node 1
NETNAME_VALUE = DB2SRV1
NETNAME_DEPENDENCY = IP Address for Node 1
#DISK_NAME = Disk O:
#TARGET_DRVMAP_DISK = Disk O:
please follow the isntallation doc it has given very clearly. page number 254.
if u need any further guidence i will.
Kiran have u finshed the installation
Regards
Rao .
Edited by: Rao on Jun 22, 2009 12:44 PM

2 CIs in a MSCS Cluster

Dear Portal Gurus,
Can we install 2 CIs in each node of a 2 node MSCS Cluster ?
(we have 2 noeds i.e. 2 hosts)
Thanks.
Jack.

Hi Jack,
No I guess. You have two options.
Install Java Central instance on 1 Node and Dialog Instance on the other one. Have SCS instance + Database on both for failover. Install other Dialog instances outside MSCS Cluster
Install 1 SCS Instance + Database on MSCS Cluster and install Java Central instance outside cluster.
Regards,
Piyush
ps: please mark for useful answers.

How to install CCMS agent in MSCS cluster system?

Hi,
We just migrated our WAS 7.0 system from a single to a MSCS cluster system and we need re-install CCMS agent on it. New cluster system have two nodes. Each node we have local SAP instances. We have central instance for message server running on a shared drive, which can failover to one of the nodes. My question is when I install CCMS agent, where I should I install it? Each node of the cluster on local drive or only shared drive or I have to install on three places?
Thanks you.
Yujun Ran

Hallo.
Did you solve your issue?
I followed Oss note 536954, but I am not sure that it is for new cluster system.
Furthermore, the CCMS agent that you installed, point to a CEN system (as Solution manager for example) and as second system to the central instance ?
Thanks.
Mario

Multiple SAP Systems In One MSCS Cluster

We have installed a NW 7.0 EHP abap stack in a MSCS Cluster with 2 nodes.
This was installed as a standard Cluster Install.
We have now higlighted a requirement to install a 2nd (Java ) instance in this Cluster Group.
The NW 7.0 EHP1 documentation and Note 1321854 shows that multiple SAP systems in one MSCS cluster is supported.
The documentation, however, is not clear if the 2nd instance can be installed when the 1st instance was installed without the preperations for Multiple Systems.
There are no further documentation or guidelines to help.
Has anyone come across this situation or managed to install 2 SAP instances in one Cluster Group based on Windows 2003, SQL Server 2005.
Regards

I am not quite sure if I understood your problem completley. But let's see...
Usually you use on on MSCS Cluster Group per SAP instance. So you have your already existing NW 7.0 ABAP stack in a separated Cluster Group, and you create a new Cluster Group for your new Java system.
You can find more information about it in this guide here on SDN:
Link: [High Availability;
More information about High Availability Setups for SAP System can be found here on SDN:
Link: [High Availability]
If you want to use the same database for your clustered ABAP stack and your new Java stack you can do a MCOD installation.
Please see Link: [SAP Note 388866 - Multiple Components on One Database - Installation|https://service.sap.com/sap/support/notes/388866]
Hope this helps.
Best regards,
Mathias

Java Add-In installation on a existing ABAP System (MSCS Cluster)

Hi all
we try to install a Java Add-In for a already existing ABAP System with
SAPinst. The ABAP System is a NetWeaver2004s SR1 System in a MSCS Cluster.
We have the Problem that we can´t choose the Installation Drive during
the SAPInst Define Parameter Step during the "Central Services Instance (SCS) Java Add-In" Installation.
But in the Installation Manual for
NW2004s Java AS in Chapter 4.3.2 is stated:
"Caution: You must install it on the same shared disk as the ASCS instance."
If I run the Installation, the SCS will be installed on the local disk E:
beside the Central Instance. But it need to be on the shared Disk P:
beside the ASCS installation. How can I control this?
Stefan

obviously the network share saploc is existing in your system an pointing to a local device.
Delete the local saploc share, install SCS instance, run MSCS first node, additional node, create the local share saploc again (don't forget to set permissions on the share, the default permission did only allow Read vor Everyone) and proceed with installation.
There is a note in creation for this issue.
peter

SAP ECC 6.0 SR3 Cluster failover not working in AIX with DB2 UDB V9.1 FP6

Hi Gurus,
We have installed the SAP ECC 6.0 SR3 High Availability with DB2 UDB V9.1 FP6 in AIX cluster environment.
After installation we are doing the cluster fail test.
Node A
Application Server
Mount Points:
/sapmnt/<SID>
/usr/sap/<SID>
/usr/sap/trans
Node B
Database Server
Mount Points:
/db2//<SID>
The procedure followed to do the cluster failover:
We have down the cluster on Node A and all the resources of the Node A has been moved to Node B.
On Node B when we issued a command to start the SAP. It says u201Cno start profiles foundu201D
WE have down the cluster on Node B and moved the Resource from Node B to Node A . There the db2 User IDu2019s are not available. We have crated the user Idu2019s manually on Node A. however it did not work.
Please suggest the procedure to start the sap in cluster failover.
Best Regards
Sija

Hi Sija,
Can i have detailed scenario in your cluster configuration.
Means you are saying that going to start cluster package manually, if it is right please make sure that you had the same copy of start, instance profiles of NodeA to Node B. Means you need to maintain two startup, two instance profiles for both nodes. In a normal situation it will picik the profile of node A to start databse from A node. But in a failover situation it will not pick node A profile to start, it should pick Node B s profiles.
Just make a copy from node A and change the profile name accordingly to Node b. Then try to restart.
Regards
Nick Loy

MSCS cluster installation on win + sql server 2005

Dear Guru,
I want to installing MSCS cluster installation on win 2003 ENT.with sql server 2005 .
Choosing High availability (ABAP)
Please Guide me prerequisit & partation on local & shared disk.
Its Urgent.
Personal mail ID is [email protected]
Regards
Amol kulkarni
Message was edited by:
amol kulkarni

Hi,
You know what ?
Step by step installation is different in every company because each company has a different usage of SAP software.
It is precisly your job as a SAP sys admin, to study the problem, to solve it and to write the step by step documentation.
In my company, it took me one month and my documentation is 250 pages long.
You will not get someone else doing YOUR job instead of you.
The standard SAP documenation is perfectly enough to study your particular case...
Good luck !
Olivier

Installing MSCS cluster installation

HI Guru,
I want to installing MSCS cluster installation on win 2003 ENT.with sql server 2005 .
Choosing High availability (ABAP)
Please Guide me prerequisite & partation on local & shared disk.
Its Urgent.
Personal mail ID is [email protected]
Regards
Amol kulkarni

Hi,
You know what ?
Step by step installation is different in every company because each company has a different usage of SAP software.
It is precisly your job as a SAP sys admin, to study the problem, to solve it and to write the step by step documentation.
In my company, it took me one month and my documentation is 250 pages long.
You will not get someone else doing YOUR job instead of you.
The standard SAP documenation is perfectly enough to study your particular case...
Good luck !
Olivier

Sap uninstalltion in mscs cluster

Dear All,
Can any one help me in uninstalling the sap in mscs cluster environment
windows 2003 sp2
netweaver 2004s
oracle 10g
2 node cluster
<b>
I have tried the following to resolve the issue without success
1.tried from sapinst
2.tried uninstalling it from run box with virtual host name</b>
The below is the error message i get when im trying to uninstall the system
WARNING 2007-10-31 17:08:58
Connect to message server (sapxpi/3900) failed: NIECONN_REFUSED.
INFO 2007-10-31 17:08:58
Execute step removeJavaInstancesFromDB of component |NW_Uninstall|ind|ind|ind|ind|0|0|NW_System_Uninstall|ind|ind|ind|ind|3|0|NW_Instance_Uninstall|ind|ind|ind|ind|2|0.
INFO 2007-10-31 17:08:58
Execute step runCleanIPC of component |NW_Uninstall|ind|ind|ind|ind|0|0|NW_System_Uninstall|ind|ind|ind|ind|3|0|NW_Instance_Uninstall|ind|ind|ind|ind|2|0.
ERROR 2007-10-31 17:08:58
FJS-00003 TypeError: instance has no properties (in script NW_Uninstall|ind|ind|ind|ind, line 90104: ???)
ERROR 2007-10-31 17:08:58
FCO-00011 The step runCleanIPC with step key |NW_Uninstall|ind|ind|ind|ind|0|0|NW_System_Uninstall|ind|ind|ind|ind|3|0|NW_Instance_Uninstall|ind|ind|ind|ind|2|0|runCleanIPC was executed with status ERROR .
INFO 2007-10-31 17:09:14
An error occured and the user decided to retry the current step: "|NW_Uninstall|ind|ind|ind|ind|0|0|NW_System_Uninstall|ind|ind|ind|ind|3|0|NW_Instance_Uninstall|ind|ind|ind|ind|2|0|runCleanIPC".
ERROR 2007-10-31 17:09:15
FJS-00003 TypeError: instance has no properties (in script NW_Uninstall|ind|ind|ind|ind, line 90104: ???)
ERROR 2007-10-31 17:09:15
FCO-00011 The step runCleanIPC with step key |NW_Uninstall|ind|ind|ind|ind|0|0|NW_System_Uninstall|ind|ind|ind|ind|3|0|NW_Instance_Uninstall|ind|ind|ind|ind|2|0|runCleanIPC was executed with status ERROR .
INFO 2007-10-31 17:09:21
An error occured and the user decide to stop.\n Current step "|NW_Uninstall|ind|ind|ind|ind|0|0|NW_System_Uninstall|ind|ind|ind|ind|3|0|NW_Instance_Uninstall|ind|ind|ind|ind|2|0|runCleanIPC".

Hello,
I encountered the similar situation & got good result following Mr.Camilo Lopez de Felipe.
The command to start SAPinst was
sapints.exe SAPINST_USE_HOSTNAME=<virtual hostname for SAP>
Thank you very much.
Error Massage
INFO 2009-08-28 16:47:52
Execute step runCleanIPC of component |NW_Uninstall|ind|ind|ind|ind|0|0|NW_System_Uninstall|ind|ind|ind|ind|3|0|NW_Instance_Uninstall|ind|ind|ind|ind|2|0.
ERROR 2009-08-28 16:47:52
FJS-00003 TypeError: instance has no properties (in script NW_Uninstall|ind|ind|ind|ind, line 90104: ???)
ERROR 2009-08-28 16:47:53
FCO-00011 The step runCleanIPC with step key |NW_Uninstall|ind|ind|ind|ind|0|0|NW_System_Uninstall|ind|ind|ind|ind|3|0|NW_Instance_Uninstall|ind|ind|ind|ind|2|0|runCleanIPC was executed with status ERROR .
Edited by: Kaoru Takahata on Aug 28, 2009 12:02 PM

PI 7.1 Installation on MSCS Cluster

Hi Sapers,
                     I'm In the middle of Installation of PI 7.1 on MSCS Cluster
                     My Problem is while Installing In Node A I have chosen for Host Agent and SAPOSCOL a Shared Drive(G:) which
                     Should actually be a Local Drive(C , But In Node B I have Chosen Local Drive (C
                     Now My question
                     1) Can I change the path In the Registry, If yes can anyone tell me the procedure and
                      2) What would be the implications or Consequences?
                      3) Will My Installation Go Successfully
Thanks In Advance
GHK

Installation of SAPOSCOL on a different logical partion within the same local host is fine, obviously not on a different host as the SAPOSCOL is the operating system collector for the host on which it is installed.
Regards.

Doubt in WLS 7.0 JMS Cluster Failover

I configured 2 JMS servers JMS1 and JMS2. I also configured 2 JMS file stores FS1 and FS2. JMS server JMS1 using FS1 and JMS2 using FS2. I also configured 1 distributed queue DQ for the cluster server. I also configured 1 local queue LQ1 in JMS1 and 1 local queue LQ2 in JMS2.
          When the JMS2 server failed, I migrate it to the migratable target JMS1. The file store FS2 and all transaction log files have been copied to the JMS1 server with the same path. I first migrated JTA service and then JMS service. The migration procedure worked and no problem. But the messages in the JMS2 did not get processed by the migratable target JMS1.
          What was wrong in my configuration? In my migration procedure? Or in my file store? Or anything? I knew I should use shared disk storage for file store. I just want to simulate the problem and try cluster failover.


          Chan Tong wrote:
          > Tom Barnes <[email protected]> wrote:
          >
          >>
          >>Chan Tong wrote:
          >>
          >>>I configured 2 JMS servers JMS1 and JMS2. I also configured 2 JMS file
          >>
          >>stores FS1 and FS2. JMS server JMS1 using FS1 and JMS2 using FS2. I
          >>also configured 1 distributed queue DQ for the cluster server. I also
          >>configured 1 local queue LQ1 in JMS1 and 1 local queue LQ2 in JMS2.
          >>
          >>>When the JMS2 server failed, I migrate it to the migratable target
          >>
          >>JMS1. The file store FS2 and all transaction log files have been copied
          >>
          >>
          >>>to the JMS1 server with the same path. I first migrated JTA service
          >>
          >>and then JMS service. The migration procedure worked and no problem.
          >>
          >>>But the messages in the JMS2 did not get processed by the migratable
          >>
          >>target JMS1.
          >>
          >>>What was wrong in my configuration? In my migration procedure? Or in
          >>
          >>my file store? Or anything? I knew I should use shared disk storage
          >>for file store. I just want to simulate the problem and try cluster
          >>failover.
          >>
          >>Need more information. What do you mean by "the messages in the JMS2
          >>
          >>did not get processed"? Do JMS2's message counts
          >>show up in console monitoring?
          >>
          >
          > After migrate JMS2 to JMS1, the messages in the JMS2 file store were not processed
          > by the JMS1. Could it be I manual copied the file store? Or I will need to use
          > shared file store that can be access from from both servers?
          Either way yes. The migrated JMS server still needs to be able to
          find its store. Another way to do this is to use a JDBC store - as
          long as the remote database is highly available and your
          applications can tolerate the likely JMS performance reduction.

Oracle DB upgrade from 10.2.0.2 to 10.2.4.0 under MSCS cluster

Hi,
How to upgrade the database from 10.2.0.2 to 10.2.4.0 under MSCS cluster environment.
We are in the process of upgrading the Oracle from 10.2.0.2 to
10.2.0.4.
( 10g Relese 2 (10.2.0.4) patchset 3 for Microsoft Windows X64 -
p6810189_10204_MSWIN-x86-64 )
As per the readme.txt, we have stoped all the oracle services. We could
not stop listner - OracleP01102TNSListenerFSLoracleP01. We got the
error listner ( OracleP01102TNSListenerFSLoracleP01) need to be
shutdown. We tried to stop it, but, automatically it was restarting.
We disable it and continued the upgradation. All the oracle resources
moved to Node-2 and we got the second error ( (oracle_home)/database )
device is not ready.
Early reply is highly appreciated..
Thanks!
Nikhil

Hi ,
Shutdown Database (use sqlplus) and then try to stop all oracle services.
Check with your OS team for - All the oracle resources
moved to Node-2 and we got the second error ( (oracle_home)/database )
device is not ready.
Normally it should not be happened .
If node 1 is down , on that time node 1 all the oracle resoures moved to Node -2 .
Pls contact OS team.

RDP Services do not accept connection after cluster failover

Hi guys,
i am having weird behaviour on my Windows Server 2012 R2.
server 1
- 10.100.1.201
server 2 - 10.100.1.203
VIP - 10.100.1.202
when i perform remote desktop session to server 1 and server 2 after both servers are being rebooted, they are working perfectly fine. during the remote desktop session, i perform a cluster node failover switching node to server 2. immediately i perform
the task, my server 1 connection will hang and not able to login anymore.
strangely, when i am connection from the same server zone and perform remote desktop, they work perfectly fine and will not disconnect me from neither of the server 1 and 2.
i am suspecting the network routing mess up during the cluster failover, but from the route print, there are identical and has no problem with it.
any one here has the same problem i experience?
zhiyuan

Hi,
sorry if i lose you ... here is the story
node 1 - 10.100.1.201
node 2 - 10.100.1.203
vip - 10.100.1.202
when both server restarted, i can remote to both servers, no problem.
1. RDP to both node 1 and 2 together on their physical IP. Connected successfully.
2. checked the Active node on node 1. Perform fail over from node 1 to 2, node 1 RDP session loss connection immediately. checked on node 2, cluster node active on node 2. no errors.
3. perform node 2 to node 1 fail over. Node 2 RDP session loss connection immediately, node 1 session came back. checked cluster node active on node 1. no errors.
4. in order to have both can continue to rdp, perform restart on node 2 (the node cannot reconnect), after reboot, rdp back to normal.
5. firewall team confirm connection has reach server, server not responding to rdp apparently.

Mscs cluster failover

Similar Messages

Maybe you are looking for