Lets say company A have multiple branches with this existing wireless design:
- each branch has its own C9800 with 9K APs. the config is unified across all branches but SSID different.
> Not 100% unified but you can work with that.
- the design is FLex local switching and APs configured in a way that act as independent controller ( Authenticator ) means the controller is only needed when new AP associate, AP reloads or any changes to apply on APs other than that shutting WLC will not affect the wireless clients ( existing and new ).
> What do you mean, the controller needs to stay online in local and or FlexConnect mode, an outage is okay, not designed to keep it off.
- two type of SSID there, PSK and dot1.x and for dot1x ISE is the radius
> I'm assuming ISE is in the DC?
- each site has single standalone C9800 appliance
> This is one of your biggest issues along with the different SSID's but again, there are some work adounds.
note: for the existing design since all major services are remote at DC, and Flex Local SW then I would do a central HA pair of Controller rather than having standalone at each site ( but controllers were already purchased and I could not make it happen )
> Let's ignore this then.
NOW SEEKING High Availability design recommendation:
- get a secondary node for each site and setup HA ? I personally think based on the existing design this is not a best choice also the cost is high
>
ets say company A have multiple branches with this existing wireless design:
- each branch has its own C9800 with 9K APs. the config is unified across all branches but SSID different.
- the design is FLex local switching and APs configured in a way that act as independent controller ( Authenticator ) means the controller is only needed when new AP associate, AP reloads or any changes to apply on APs other than that shutting WLC will not affect the wireless clients ( existing and new ).
- two type of SSID there, PSK and dot1.x and for dot1x ISE is the radius
- each site has single standalone C9800 appliance
note: for the existing design since all major services are remote at DC, and Flex Local SW then I would do a central HA pair of COntroller rather than having standalone at each site ( but controllers were already purchased and I could not make it happen )
NOW SEEKING High Availability design recommendation:
- get a secondary node for each site and setup HA ? I personally think based on the existing design this is not a best choice also the cost is high
> It is a higher cost, but the only way is to really have two controllers in the DC since all your ap's are in FlexConnect mode. You can probably migrate one at a time or buy a controller in the DC, bring that up, move site A ap's to that and then move site A controller to the DC. Now, the question is, are they different models?
- setup an HA pair at DCenter and point secondary controller (Backup primary controller ) on all APs at each site to it, so in case of primary site WLC failure then the APs will associate to the DC controller? and fallback when primary comes back online I am thinking this is more reasonable option.
>
ets say company A have multiple branches with this existing wireless design:
- each branch has its own C9800 with 9K APs. the config is unified across all branches but SSID different.
- the design is FLex local switching and APs configured in a way that act as independent controller ( Authenticator ) means the controller is only needed when new AP associate, AP reloads or any changes to apply on APs other than that shutting WLC will not affect the wireless clients ( existing and new ).
- two type of SSID there, PSK and dot1.x and for dot1x ISE is the radius
- each site has single standalone C9800 appliance
note: for the existing design since all major services are remote at DC, and Flex Local SW then I would do a central HA pair of COntroller rather than having standalone at each site ( but controllers were already purchased and I could not make it happen )
NOW SEEKING High Availability design recommendation:
- get a secondary node for each site and setup HA ? I personally think based on the existing design this is not a best choice also the cost is high
- setup an HA pair at DCenter and point secondary controller (Backup primary controller ) on all APs at each site to it, so in case of primary site WLC failure then the APs will associate to the DC controller? and fallback when primary comes back online I am thinking this is more reasonable option.
> This is fine, you just need to think it out a bit.
- leave it as it is, and if Controller fails then we have 4hrs RMA and the odds of having APs power outage is low ?
> What is the customer requirements?
- any other option?
> I would try to get two controllers in the DC N+1 and use that for all sites.
- leave it as it is, and if Controller fails then we have 4hrs RMA and the odds of having APs power outage is low ?
- any other option?
- setup an HA pair at DCenter and point secondary controller (Backup primary controller ) on all APs at each site to it, so in case of primary site WLC failure then the APs will associate to the DC controller? and fallback when primary comes back online I am thinking this is more reasonable option.
- leave it as it is, and if Controller fails then we have 4hrs RMA and the odds of having APs power outage is low ?
> Its up to the customer requirements.
- any other option?
-Scott
*** Please rate helpful posts ***
