Forum Discussion
Network Failover
- Mar 26, 2016
Yes, and it's generally recommended to do so.
I'd even recommend to use two physical interfaces (as Aggregated Trunk) for HA, unless you plan to use all other 7 interfaces for something else. It would make sense to use that HA trunk interface for three things: Config Sync, Network fail-over, Traffic Mirror
Yes, and it's generally recommended to do so.
I'd even recommend to use two physical interfaces (as Aggregated Trunk) for HA, unless you plan to use all other 7 interfaces for something else. It would make sense to use that HA trunk interface for three things: Config Sync, Network fail-over, Traffic Mirror
- MM_F_147944Mar 28, 2016
Nimbostratus
thanks for your reply kindly guide me If we have HA both from Hard Wire and Network configured and we have configured Vlan Failsafe, dose failover happens if we shut the Vlan of Active F5, however it can detect heartbeat on Hardwire, - natheMar 28, 2016
Cirrocumulus
In normal circumstances HW and network failover together the HW failover takes precedence. However, in your case vlan failsafe has Reboot as the default action. So disabling this vlan should cause the active to reboot and failover to occur. That's my view anyway. Note, I wouldn't myself specify vlan failsafe in the HA vlan. Interested to hear other views on that. - Hannes_RappMar 28, 2016
Nimbostratus
Your active appliance should reboot(or take another chosen action) 90 seconds after disabling a VLAN where the VLAN failsafe feature is enabled, regardless of hardwired failover that is in place. If you use VLAN failsafe feature, the HA VLAN itself doesn't have to be monitored. I do not see a good reason to enable VLAN failsafe feature on HA VLAN. Not to say that it's a best practice or anything, but my personal preference is to never use hardwired failover. It's great due to it's responsiveness, but it can only check the appliance's health on a very low level. I find the network failover feature reliable, and what's most important, it doesn't have any limitations (more on limitations in SOL article). Another option is to have both in place, hardwired failover as well as network failover, but I would rather keep things simple and stick to just one. You can read more on topic here: https://support.f5.com/kb/en-us/solutions/public/2000/300/sol2397.html - MM_F_147944Mar 29, 2016
Nimbostratus
thanks Nathan and Hannes, Dear Hannes I am not talking about HA Failsafe I am talking about Data Valn Failsafe , that if we Shut our Data Vlan will the status of the device will be change to Standby and peer will become active however heartbeat is still receiving on Hardwire - Hannes_RappMar 29, 2016
Nimbostratus
That VLAN failsafe@HA VLAN input was a response to Nathan's comment. I agree with him on that it would not make sense. It does not matter if we're talking about VLAN failsafe on a data VLAN, or the HA VLAN - everything I said still applies and a failover will occur. A heartbeat response via the hardwire failover cable does not prevent a VLAN failsafe action from triggering when conditions are met. - MM_F_147944Sep 21, 2016
Nimbostratus
What if on Internal Vlan Fail-over is enabled and HA link is disconnected will both the units Act as Active Active, Active Standby or Standalone ?
- Hannes_RappSep 21, 2016
Nimbostratus
What if on Internal Vlan Fail-over is enabled and HA link is disconnected will both the units Act as Active Active, Active Standby or Standalone ?
If you disconnect HA link, both units will switch to Active (regardless if VLAN Fail-over is used or not).
- MM_F_147944Sep 21, 2016
Nimbostratus
If both Units become Active what about the Sessions how it handle the Sessions a loop will create, or traffic will be confused and no sessions will create? As Switches on which Units are connected are Nexus and Port channel is created.
- IainThomson85_1Sep 21, 2016
Cumulonimbus
Lets just avoid that scenario.... All manner of badness.
You'll likely get layer 2 issues on your attached switch, and all manner of craziness on the backend.
- Hannes_RappSep 21, 2016
Nimbostratus
This is when you will have an incident. Eventually there will be ARP conflicts as both units will compete by announcing their MAC addresses as the owner of Floating SelfIP addresses. This could cause asymmetric routing problems.
Recent Discussions
Related Content
* Getting Started on DevCentral
* Community Guidelines
* Community Terms of Use / EULA
* Community Ranking Explained
* Community Resources
* Contact the DevCentral Team
* Update MFA on account.f5.com