I have a situation where having two different SAP servers with same hostname but different IP address. One located in Production Data Center and the other on DR. The reason why the server in DR has same hostname as in the Production Data Center is to have an easy setup from the application side for the DR purpose (according to the SAP team). My questions:
1. What would be the best practice for this when restoring database to the DR location? So far I have created an alias name for the server located in DR when registering in the Networker client resources. Add the alias hostname in the the /etc/hosts of the networker master server & client. It's been working so far when restoring from the server located in DR, although I have seen a warning message that the server located in Data Production Center (let's name it abc) conflicting with server located in DR (let's nameit abcDR) in the NSR database.
2. When I tried to install and configure the abcDR server as storage node for the SN managing device for clone, I got an error like this :
Call to lg_getaddrinfo() failed with host: Error 0
05/20/18 22:54:42.531047 RPC Authentication: Client failed to obtain RPCSEC_GSS credentials: Authentication error; why = Server rejected credential
05/20/18 22:54:42.531138 Could not get a session key for GSS authentication. Perhaps this authentication method is not allowed/supported by both the local and remote remote machines.
05/20/18 22:54:42.531787 Failed to create lgto_auth credential for RPCSEC_GSS: Could not get session key from abc for GSS authentication with abcDR: Authentication error; why = Server rejected credential
05/20/18 22:54:42.531874 Received mmd binding information: mmd host abcDR, prog 390104, vers 205
05/20/18 22:54:42.531906 Connecting to nsrmmd on host abcDR (prog 390104, vers 205)...
05/20/18 22:54:42.532005 DNS lookup failed for abcDR: node name or service name not known
05/20/18 22:54:42.532249 Failed to connect to mmd (prog 390104, vers 205) on `abcDR': Failed to connect to host 'abcDR': Unknown host
98519:nsrrecopy: Unable to setup direct save with server tbssvr-emcbck: Failed to connect to storage node on abcDR: Failed to connect to host 'abcDR': Unknown host.
What would be the best practice so I can use this abcDR as a destination storage node managing the clone pool?
- device using ddboost (DFC)
- clone figure :
rd=abc:dddc_ddboost_device1 -> rd=abcDR:dddrc_ddboost_device1
Appreciate for any inputs and comments.
You have made NetWorker's life more complex.
Please let us know what is the destination you are using to store this SAP server's data? DD or AFTD?
If I get this right, ABC is your DC Client and DRABC is the machine on DR.
Create etc entry on Backup server for DRABC only when you plan to recover data for this machine. Why do you want to change the SN and point it to DRABC. Keep it as it is and let the data flow from DD/AFTD to DRABC client.
Thanks for your respond. I will not create complexity if hostname of the server in the DR can be renamed. Unfortunately SAP team won't allow us.
>>Please let us know what is the destination you are using to store this SAP server's data? DD or AFTD?
As mentioned earlier, we use Data Domain DDBoost
>>If I get this right, ABC is your DC Client and DRABC is the machine on DR.
both hostname are ABC. But in the Networker Client resource, yes, I use ABC & DRABC as the name. But nsr peer information shows that only ABC exist.
>>Create etc entry on Backup server for DRABC only when you plan to recover data for this machine. Why do you want to change the SN and point it to DRABC. Keep it as it is and let the data flow from DD/AFTD to DRABC client.
I have no issue in restoring so far. But restoring quite challenging as I use DFC (DDBoost over FC). Currently I use a backup volume that connected to DFC and clone to a volume that connect to DDBoost over LAN. (e.g. rd=abc:dddc_ddboost_device1 -> rd=xyz:dddrc_ddboost_device1). When restoring SAP, I mount the clone volume to a DFC device through DRABC (DRABC act as a SN). By product design Networker will find the DDBoost over LAN connection when restoring from clone volume (in my case) and we have to unmount this same volume (the clone volume that connects to the DDBoost over LAN). The challenge is, many times I have to stop the ongoing clone process to perform restore on the DR (I need to unmount the DDBoost over LAN on the DR).
So if I can do DFC to DFC device cloning (specifically uses abc and abcDR as the SN), I believe I can do both clone and restore at the same time without having to unmount the volume that connects to DDBoost over LAN connection.
My logic is, if abcDR can be recognized/registered as abcDR in the NSR DB, then I think it's possible to do this.
I am not sure how much fruitful this might turn out but you can give this a try, Segregate your clone for DDBoost over LAN by introducing a new Network interface on abcDB and then add it as a client which will only handle clone jobs, this will help you avoid the task of unmounting LAN interface.