Unsolved
This post is more than 5 years old
24 Posts
0
1046
December 3rd, 2008 04:00
Using SRDF for DR of ECC 6.1
Hi,
I'm going to be putting in an ECC 6.1 implementation and need to consider DR.
We will have ServerA as the ECC Server/Repository and ServerB as the DR server.
Both servers boot from local disk and therefore have unique hostnames within the active directory.
For this reason we have had a DNS entry created called ECCSERV.ourcompany.com , this is the name we want all the master agents in our environment to use for communication.
In the event of a DR, we would failover the SRDF disk to site B, amend the DNS entry to target the DR server, start up ServerB and hope it all works
When installing the primary server, it's asking for primary network name and the only entry in the dropdown is ServerA.ouradname.com
Is there a way of overriding this entry at all either pre-install or post-install, if not it seems our design is useless.
I'm going to be putting in an ECC 6.1 implementation and need to consider DR.
We will have ServerA as the ECC Server/Repository and ServerB as the DR server.
Both servers boot from local disk and therefore have unique hostnames within the active directory.
For this reason we have had a DNS entry created called ECCSERV.ourcompany.com , this is the name we want all the master agents in our environment to use for communication.
In the event of a DR, we would failover the SRDF disk to site B, amend the DNS entry to target the DR server, start up ServerB and hope it all works
When installing the primary server, it's asking for primary network name and the only entry in the dropdown is ServerA.ouradname.com
Is there a way of overriding this entry at all either pre-install or post-install, if not it seems our design is useless.
No Events found!
seamuscoffey
472 Posts
0
December 3rd, 2008 06:00
Check out Patch 3928 (Backup and Recovery for Migration and Disaster Recovery (Single and Distributed Infrastructure)) on Powerlink - it's under the 6.0 downloads but the same principles apply for 6.1.
It allows for the disaster recovery of an infrastructure server to new hardware.
Regards,
Séamus Coffey
EMC Global Services
Martyn1
24 Posts
0
December 3rd, 2008 06:00
I'd already read that document, it's pre-req 3 that causes us an issue -
'the original and new hosts must retain the same fully qualified network name'
As these are 2 Windows servers that are in an Active Directory config, the names are not the same.
karan2
1 Rookie
•
32 Posts
0
December 3rd, 2008 06:00
Its just not the server DNS that you need to failover but also the registry entries of the ServerA. ECC DR can be more easily accomplished if the servers are boot from SAN.
About changing the ECC server name, I am not sure we can do this, since its embedded into the ECC repository.
Thanks,
Karan
seamuscoffey
472 Posts
0
December 3rd, 2008 07:00
Hostname changes are not supported in ControlCenter. Changing the relevant ini files alone is not sufficient as there are too many references to the original hostname embedded in the repository.
Regards,
Séamus
mamer1
18 Posts
0
August 17th, 2011 07:00
We are also struggling with this issue. We have a distributed environment. HostA is running the Server/Store/Repository and HostB is running StorageScope/Performance Manager. I have read the readme of patch 4938 and what I'm reading isn't pleasing. First of all, it is a step by step manual in the assumption the HostA is up and running. I am wondering what happens when you do not have the ability anymore to contact HostA. You do have an SRDF relationship and both local and remote BCVs. And you never ran the backup util mentioned in the readme.
Other questions are:
- What does backup.bat backup exactly? I tried to read the Perl script, but I am a Storage Administrator not a Perl programmer.
- Why can't I find any other white papers or documentation about this fairly important issue.
- Why is the embedded oracle database so depended on the original hostname?