Need to Understand the Hardware (Serial) Failover Vs Network Failover in F5 - LTM v11.3?

See this article for a good description of the two.

http://support.f5.com/kb/en-us/solutions/public/2000/300/sol2397.html?sr=34012910

Pay attention to the portion that states if you use both, the hardwire supercedes the network.

Inspite of the recomendations, I have been using network failover exclusively since 9.4. Initially this was based on distance, but more recently I have just come to trust it.

Hi,

Thanks for the article to refer.

But, i want to understand that which features or benefits we have in network Failover to use?

As they mention in the articles; "Important: Even if you configure hardwired failover, communication over the network is necessary for certain features to function properly, for example, the communication that occurs over the network during failover mirroring."

Can you please explain it briefly?

Two features that I believe require network failover are connection and persistence mirroring.

http://support.f5.com/kb/en-us/solutions/public/13000/400/sol13478.html?sr=34013210

I'm not sure which other features require it.

Thanks for it.

But, where it is mentioned that they require a network Fail-over? I read the article, but couldn't find it. Can't these two feature can be achieve in Hardware Fail-over.

Please explain.

The features require a network connection between the devices, if that can be a non-failover connection, I do not know.

Going back to the first link I posted, which was updated last month, hardware failover is still recomended due to faster response times. Under that where it mentions certain features need a network connection, it states failover mirroring, which I believe are the features described in the second link I sent.

Beyond that, I do not have any information. I went through this same conversation last year when for the first time I had BigIPs close enough together to use the hardware failover. Due to the fact I had been using network failover for many years, and that some hardware lacks the serial port to handle hardware failover, my account team and I decided my best choice was to stay with network failover.

Given your situation is the exact opposite of mine, I would question whoever made the recomendation for you to switch to network and ask them why.

Thanks a lot.

Appreciated!!!

Hard-wired failover When you configure hard-wired failover, you enable failover by using a failover cable to physically connect the two redundant units. This is the default setting.

Network failover When you configure network failover, you enable failover by configuring your redundant system to use the network to determine the status of the active unit. You can use network failover in addition to, or instead of, hard-wired failover. When using network failover, you can configure the system to use either self IP addresses or management IP addresses for failover. For more information on specifying IP addresses for failover

By default, the way that a BIG-IP unit monitors the status of its peer, in order to detect that failover is required, is through a hard-wired connection between the two BIG-IP units. With proper configuration, however, you can cause each BIG-IP unit to monitor peer status by way of a TCP/IP network connection instead. For more information on types of failover connections, see Configuring the failover type.

Configuring High availability - http://support.f5.com/kb/en-us/products/big-ip_ltm/manuals/product/bigip_nsm_guide_943/nsm_high_avail.html

Thats a nice article... Thansks :)

Can we list down that what hardware fail-over can do or not and same for the Network fail-over? in order to understand.

Both have the same purpose to switch the Active unit in case of failure, Hardware fail-over will take action when one of the hardware fails where as the Network Fail-over comes into play if the particular network link fails.

Both have the same purpose to switch the Active unit in case of failure, Hardware fail-over will take action when one of the hardware fails where as the Network Fail-over comes into play if the particular network link fails.