Network Failover Steps
Hello,
From a network administrator's perspective, what steps should be taken to either test a failover or go through one actually?
For the sake of simplicity, we'll consider an HO and a DR location equipped with the same equipment. The DR is hot with data being replicated. I've attached an ascii representation of the network.
What step by step procedures must be in place to expedite a smooth failover?
You don't have to write it up yourselves, just point me to the links or documents that'll help me get the answers.
Thanks for looking.
ASCII-Network.JPG
From a network administrator's perspective, what steps should be taken to either test a failover or go through one actually?
For the sake of simplicity, we'll consider an HO and a DR location equipped with the same equipment. The DR is hot with data being replicated. I've attached an ascii representation of the network.
What step by step procedures must be in place to expedite a smooth failover?
You don't have to write it up yourselves, just point me to the links or documents that'll help me get the answers.
Thanks for looking.
ASCII-Network.JPG
ASKER
I'm willing to open another ticket for those on the fence - points-wise. I need help on this - bad!
Thanks
Thanks
>From a network administrator's perspective, what steps should be taken to either test a failover or go through one actually?
Typically through a maintenace window, there is no such thing as a smooth failover (Typically) at layer 2 or 3 depending on the protocols that you are using. The process is stright forward, just start pulling cables and monitoring the failover.
>Cases, not complete by any means, but a start:
All your cases are strong and are assumptions based on the configuration of the equipement and protocols, so with that being said, start testing at the ISP level by disconnecting one cable and testing failover. But, honestly, a realistic recommendation can not be given based on what you have without additional information (Strong Network Diagram, configs, etc).
Billy
Typically through a maintenace window, there is no such thing as a smooth failover (Typically) at layer 2 or 3 depending on the protocols that you are using. The process is stright forward, just start pulling cables and monitoring the failover.
>Cases, not complete by any means, but a start:
All your cases are strong and are assumptions based on the configuration of the equipement and protocols, so with that being said, start testing at the ISP level by disconnecting one cable and testing failover. But, honestly, a realistic recommendation can not be given based on what you have without additional information (Strong Network Diagram, configs, etc).
Billy
ASKER
I was hoping for a general procedure, not necessarily a custom made one to suit me.
Pulling cables or turning off the ports would be the ideal way to test the failover theory. Of course, dynamic routing at the core helps. Vlans trunked across the link to the DR helps keep the segments segregated.
I know all this. But, how does one go about documenting it? What procedures need to be implemented?
Thanks Billy.
Pulling cables or turning off the ports would be the ideal way to test the failover theory. Of course, dynamic routing at the core helps. Vlans trunked across the link to the DR helps keep the segments segregated.
I know all this. But, how does one go about documenting it? What procedures need to be implemented?
Thanks Billy.
ASKER
Anyone?
>I was hoping for a general procedure, not necessarily a custom made one to suit me.
That is the idea, there is not a general procudure that fits all; depending on the protocols that you have used in your network will dictate on how the network will fail-over and reconverge. Once you understand your network, you can build out the procedure. If you understand the protocols and traffic flows, etc, you will understand how the failover will occur; this is also usually a step in the design process.
Billy
That is the idea, there is not a general procudure that fits all; depending on the protocols that you have used in your network will dictate on how the network will fail-over and reconverge. Once you understand your network, you can build out the procedure. If you understand the protocols and traffic flows, etc, you will understand how the failover will occur; this is also usually a step in the design process.
Billy
ASKER
Billy,
Can you give me something more concrete?
Steps to test routing, verify connectivity, change configs - anything?
Can you give me something more concrete?
Steps to test routing, verify connectivity, change configs - anything?
ASKER CERTIFIED SOLUTION
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
ASKER
1. If the internet connection at the active fails, the internet connection at the standby should take over and service requests, both internal and external.
2. If the internet router at the active fails, the internet router at the standby should take over and service requests, both internal and external.
3. If the ASA at the active site fails, the ASA at the standby site should take over and service requests.
4. If the ISA at the active site fails, the ISA at the standby should take over and service requests.
5. If the connection between the active and standby sites fails, branches should be able to communicate via the MPLS to the active site.
6. If the MPLS router at the active site fails, the MPLS router at the standby site should be able to route traffic from the branches to the active site.
I can't figure out exactly what steps to take or the procedure to test the failover.
Any and all input would be highly appreciated.