Oracle data guard for Distributed SAP system

[Former Member]

Hi all,

During setup of oracle dataguard configuration between 2 SAP systems running with oracle , usually tnnames.ora file is adjusted with PRIM and STBY locations an updated to

/oracle/SID/102_64/network/admin and /sapmnt/SID/profile/oracle

I think updating in to above 2 locations we could tnsping and successfuly communicate system where CI + DB are running on one host in PRIMARY and STANBY location.....this is sandard scenario...

_________________________________________________________________________________________

Now i have PRIMARY SAP system as distributed , CI running on SYSTEMA and DB running on SYSTEMB.

If i configure tnsnames.ora file pointing between SYSTEMB and DRSYSTEM,

can i upload same file tnsnames.ora in /sapmnt/SID/profile/oracle...even application is on other host or

Do i need to configure file in tnsnames.ora pointing to SYSTEMA and DRSYSTEM..

In target standby location there is only one HOST where (CI+B) is running -- DRSYSTEM

Hope i am clear with my inputs, Please advice.

..

thanks,

rahul

[Former Member]

default gateway is not maintained in STANDBY system, therefore ping failed vice versa.

I have configured DR ebtween 2 systems and hit another issue. I have opened a new post can you kndly look look in to the issue..

http://forums.sdn.sap.com/post!post.jspa?forumID=220&reply=false

SQL> select status, Error from v$archive_dest where dest_id=2;

STATUS ERROR

-

---

-

---

ERROR ORA-00270: error creating archiv

[Former Member]

Need to understand ping status in clustered environment in DR setup's

[Former Member]

Hi,

As I understand that you tried to ping DR (192.168.30.23) on database host 192.168.10.35 and failed. At this stage, your assumption is correct. You should be able to ping the DR. Check your network configuration.

Best regards,

Orkun Gedik

[Former Member]

Thanks Gedik,very clear and confident this is going to work out.

[Former Member]

Hi Gedik,

Following are my ips...

Ci+DB

192.168.30.23 eccprddr - DR system

HA

192.168.10.35 eccdb eccdb.india123.com - virtual node

192.168.10.37 eccacs eccacs.india123.com - virtual node

192.168.10.25 xyzdb - physical nodes

192.168.10.27 abcacs eccacs - physical node

Configured DR between

192.168.30.23 eccprddr - DR system

and

192.168.10.35 eccdb eccdb.india123.com - virtual DB node on which is on physical NODE - 192.168.10.25 xyzdb

I configured /etc/hosts file beteween 2 above DR systems and tried to ping but the following is result..which is not normal..

xyzdb:/#ping 192.168.30.23

PING 192.168.30.23: 64 byte packets

-

---

192.168.30.23 PING Statistics----

65 packets transmitted, 0 packets received, 100% packet loss

--

I believe unkess ping from osurce and destitnation hapeens fine the log shipping will not happen, pl let mw know if this is the case.

thanks,

ravi

[Former Member]

Hi Rahul,

At the DR system side, primary system virtual hostname must be configured in the tnsnames.ora. This is because, when nodeA is failed, nodeB will take all the resources on itself. By doing so, DR system will continue to run, even if the nodeA is failed. This is DR to primary scenario.

On the other hand, with the primary to DR scenario, physical DR hostname must be configured in the tnsnames.ora. This is because, "/sapmnt" will be configured as a cluster resource and when the nodeA is failed, it will be transferred to nodeB automatically. So, the system will run on the same configuration files.

I hope that it is answered your question,

Best regards,

Orkun Gedik

[Former Member]

Hi Gedik,

I could understand to some extent, my primary is running on 2 hosts as mentioned...CI on NODE A and DB on NODE B

1) PRIMARY

I will configure tnsnames.ora pointing to DRSYSTEM from NODE B.(DB INSTANCE) .......update same in /oracle/SID/102_64/network/admin on NODE B

and update same file tnsnames.ora in /sapmnt/SID/profile/oracle/......in NODE A

2) and vice versa in STANDBY location

tnsnames.ora will be configured in DRSYSTEM pinting to NODE B.. (DB INSTANCE OF PRIMARY)

Now if i login to NODE A ( CI INSTANCE ) of primary and try tnsping sid_stby.world will this work.....even though /sapmnt/SID/profile/oralce/tnsnames.ora is only configured between NODE B and DRSYSTEM.

I am just trying to understand...

thanks,

ravi

[Former Member]

Hi,

>> 1) PRIMARY

>> I will configure tnsnames.ora pointing to DRSYSTEM from NODE B.(DB INSTANCE) .......update same in /oracle/SID/102_64/network/admin on NODE B

>> and update same file tnsnames.ora in /sapmnt/SID/profile/oracle/......in NODE A

It is correct. You should configure physical DR hostname, on both nodes

>> 2) and vice versa in STANDBY location

>> tnsnames.ora will be configured in DRSYSTEM pinting to NODE B.. (DB INSTANCE OF PRIMARY)

You should use primary site db instance "virtual hostname", not the physical on the DR configuration. Under this circumstance, because of this hostname is virtual, even if nodeB is failed, nodeA will take all the resources, including virtual database server hostname. Then, standby will continue to run without a problem. This is because DR will access to the primary by using virtual hostname.

>> Now if i login to NODE A ( CI INSTANCE ) of primary and try tnsping sid_stby.world will this work.....even though /sapmnt/SID/profile/oralce/tnsnames.ora is only configured between NODE B and DRSYSTEM.

Yes, this is because, both nodes must have same tns name for DR system. Please note that /sapmnt is a cluster resource. So, it will be move to the another instance, if any failover occurs.

I hope that it is clear, now.

Best regards,

Orkun Gedik

Edited by: Orkun Gedik on Sep 19, 2011 3:30 PM