Home » Server Options » RAC & Failsafe » Should I configure the RAC as strange topology? (Oracle 11gr2, Sun Solaris)
Should I configure the RAC as strange topology? [message #660213] Sun, 12 February 2017 09:17 Go to next message
trantuananh24hg
Messages: 724
Registered: January 2007
Location: Ha Noi, Viet Nam
Senior Member
Good day, my Orafaq!

I have a question on a case from my remote size DBA, this made me wondering.

We have a RAC with 2 nodes, connected from server directly to storage using HBA - QLogic, 8gb/s. Now, on the remote size (2000 km distance), we have 2 servers, all of them are using Solaris 11gr2, Sparc. The question is: Can I configure those remained servers into our exits RAC right now? Of course, those servers will use network - my WAN, 100 gigabits/s, instead of HBA.

I'd never ever got the strange case.

Thank you very much!

[Updated on: Sun, 12 February 2017 09:37]

Report message to a moderator

Re: Should I configure the RAC as strange topology? [message #660219 is a reply to message #660213] Sun, 12 February 2017 10:09 Go to previous messageGo to next message
Michel Cadot
Messages: 65964
Registered: March 2007
Location: Nanterre, France, http://...
Senior Member
Account Moderator

2000km means at least 10ms on fiber channel to only transfer a command.
Do you think you can have a global SGA with this latency (without speaking about throughput)?

The best you can do is a (RAC) standby database at remote site.

Re: Should I configure the RAC as strange topology? [message #660222 is a reply to message #660213] Sun, 12 February 2017 11:45 Go to previous messageGo to next message
John Watson
Messages: 7664
Registered: January 2010
Location: Global Village
Senior Member
THere was a white paper back in release 11.something that I think claimed 300km as the theoretical limit for a stretched cluster, and reference sites of up to 100km.
Re: Should I configure the RAC as strange topology? [message #660224 is a reply to message #660222] Sun, 12 February 2017 22:03 Go to previous message
trantuananh24hg
Messages: 724
Registered: January 2007
Location: Ha Noi, Viet Nam
Senior Member
Thank you, Michel and John.

I know and understand about the latency and lack of transportable package as far as possible. But this is a strange and quite responsible question from my remote size DBA.

How about the Golden Gate with long distance?

Thanks again!

[Updated on: Sun, 12 February 2017 23:00]

Report message to a moderator

Previous Topic: Oracle 12.1.0.2.0 GI Installation for Cluster fails on 2nd Node
Next Topic: Interoperating different Windows releases in one cluster
Goto Forum:
  


Current Time: Thu Nov 15 10:30:45 CST 2018