We do not have enough APs to say one way or another, but I did get a call from Michigan State saying they were having problems with controller response at about 3,000 APs
------------------------ Walter Reynolds Principal Systems Security Development Engineer Information and Technology Services University of Michigan (734) 615-9438 On Thu, Feb 19, 2015 at 3:16 PM, Kitri Waterman <[email protected]> wrote: > Matt, > > You mentioned "10 min for full HA convergence" which seems long. But then > again, we've also seen a consistent issue where during an HA event the > configs are reported as "out of sync" and the secondary controller has to > reboot, sometimes more than once to re-sync. > > My current theory is that very dynamic parts of the configs like rogue > adhoc alerts should be left out of the controller XML sync check and this > would enable faster HA pairing. > > > Kitri Waterman > -- > Network Engineer (Wireless) > University of Oregon > > > > > > On 2/19/15 11:58 AM, Williams, Matthew wrote: > > So far for us, I the benefits outweigh the risks. In our case, when we > failed one route engine the 8510s simply failed over to the HA box and > there was no perceivable degradation in the user experience. Our issues > with using Layer-2 for the HA only occurred if BOTH of our upstream > switches had a routing engine failure within the 10 minutes it takes for > the full HA convergence time. The odds of that happening at the same time > are extremely remote and would most likely indicate a much larger > networking issue that would trump a wireless outage. > > > > We chose to do the direct connection just to be extra safe. In fact, the > only reason we even found out about the routing engine issue was because we > replaced routing engines in each of our upstream switches at roughly the > same time. > > > > I don’t know how much testing you have done with the 8510s but we’ve run > them through some pretty good paces. > > > > Respectfully, > > > > Matthew Williams > > IT Manager, Wireless > > Kent State University > > Office: (330) 672-7246 > > Mobile: (330) 469-0445 > > > > *From:* The EDUCAUSE Wireless Issues Constituent Group Listserv [ > mailto:[email protected] > <[email protected]>] *On Behalf Of *Lee H Badman > *Sent:* Thursday, February 19, 2015 2:49 PM > *To:* [email protected] > *Subject:* Re: [WIRELESS-LAN] Looking for testimonials- 8510 vs 5508 WLC > reliability > > > > I saw the bowing thing- that’s insane. It makes me curious about the racks > they are mounted in. > > > > To date, do you see the advantages of the 8510 as outweighing the concerns? > > > > > > *From:* The EDUCAUSE Wireless Issues Constituent Group Listserv [ > mailto:[email protected] > <[email protected]>] *On Behalf Of *Williams, Matthew > *Sent:* Thursday, February 19, 2015 2:46 PM > *To:* [email protected] > *Subject:* Re: [WIRELESS-LAN] Looking for testimonials- 8510 vs 5508 WLC > reliability > > > > Lee, > > We’re going through this as we speak. We’ve seen some issues with the > controllers physically bowing. We’ve also had issues with running HA via > Layer-2 when upstream routing engines/supervisor cards fail. Part of the > issue is that there is very limited keep alive timer control in our version > of code and the thresholds aren’t high enough to mitigate the route engines > switching over. Cisco recommends direct connecting them, so we’ve resorted > to using switches as media converters for the direct connection. > > > > The good news is that 8.0.100 reintroduced timer controls that did > mitigate the failures when using Layer2 for the HA. The bad news is that > 8.0.100 caused watchdog failures in the 8510s. > > > > Respectfully, > > > > Matthew Williams > > IT Manager, Wireless > > Kent State University > > Office: (330) 672-7246 > > Mobile: (330) 469-0445 > > > > *From:* The EDUCAUSE Wireless Issues Constituent Group Listserv [ > mailto:[email protected] > <[email protected]>] *On Behalf Of *Lee H Badman > *Sent:* Thursday, February 19, 2015 2:35 PM > *To:* [email protected] > *Subject:* [WIRELESS-LAN] Looking for testimonials- 8510 vs 5508 WLC > reliability > > > > One more for the group: Having a large number of 5508s, I have an interest > in greatly simplifying the environment by moving to higher capacity 8510s > (zero interest in 5760s). > > > > I’m hoping to hear from those that have taken the 8510 plunge. For us, > outside of code issues, the 5508s have been solid hardware for us. Would > like to know whether current 8510 users find them to be reliable as > hardware, and whether the HA promises live up to the glossy. > > > > Thanks- > > > > Lee B. > > > > Lee Badman > > Wireless/Network Architect > > ITS, Syracuse University > > 315.443.3003 > > (Blog: http://wirednot.wordpress.com) > > > > > > > > ********** Participation and subscription information for this EDUCAUSE > Constituent Group discussion list can be found at > http://www.educause.edu/groups/. > > ********** Participation and subscription information for this EDUCAUSE > Constituent Group discussion list can be found at > http://www.educause.edu/groups/. > > ********** Participation and subscription information for this EDUCAUSE > Constituent Group discussion list can be found at > http://www.educause.edu/groups/. > ********** Participation and subscription information for this EDUCAUSE > Constituent Group discussion list can be found at > http://www.educause.edu/groups/. > > > ********** Participation and subscription information for this EDUCAUSE > Constituent Group discussion list can be found at > http://www.educause.edu/groups/. > > ********** Participation and subscription information for this EDUCAUSE Constituent Group discussion list can be found at http://www.educause.edu/groups/.
