>Do you just get user complaints when no one can connect, or is >there some way to automatically monitor the dashboard, and reboot >APs that have had zero clients for a long time? (Is there a report >that gives daily users per AP? I have not found anything better than >checking the APs list several times a day, though it would be handy >to see peak client counts and such.) Is there any way to proactively >reboot groups of APs early in the morning or something?
Great question. We do get legitimate complaints...but they are VERY difficult to identify amid all of the "wifi sucks <http://wirednot.wordpress.com/2013/02/18/when-good-wireless-feels-bad/>" comments. You can sort APs by client count for the last 2 hours. Any AP in a residence with 0 clients in two hours gets a reboot. Note well that "0" doesn't always mean 0. We often see APs with "0" clients in the last two hours that have active clients. Meraki support can do a mass reboot remotely but that is not a user-accessible feature. I'm OK with that. If I wanted lots of knobs and levers I'd buy something else. Rand Rand P. Hall Director, Network Services askIT! Merrimack College 978-837-3532 [email protected] If I had an hour to save the world, I would spend 55 minutes defining the problem and five minutes finding solutions. – Einstein On Fri, Nov 7, 2014 at 8:53 PM, Steve Bohrer <[email protected]> wrote: > Hi Rand, > > Thanks for your comments to the wifi list about Meraki’s lack of > transparency. > > > On Oct 27, 2014, at 2:47 PM, Hall, Rand <[email protected]> wrote: > > […] > > We recently blindly updated the firmware before school started and appear > to be paying a price. That price is at least two-fold, APs regularly stop > accepting connections until rebooted (not cool) and reverting to previous > firmware doesn't seem to be an option. Engineering is on the case, but > tight-lipped. > > > […] > > We've also had a few MR16s start failing recently. I don't expect them to > last forever but hope this is not a trend. > > > We switched to Meraki from an aged Cisco WLC/WCS system last January, and > now have 110 MR16 APs. We are only about 1/8th your size in terms of > student population, though our campus is fairly spread out, so I probably > need about another 25 APs to cover most of it with reasonable density. I > had four or five AP failures within the first couple months of deployment. > Our symptoms were the AP would be completely non-responsive, with only the > power LED glowing orange, no other lights or activity. After four of these > failures, they started offering me MR34s as replacements, which was a > pretty sweet deal. Then, after seven such failures, they released a > firmware update in early June that seems to have stopped these particular > failures for us — I’ve had one MR16 die since then, but it was just totally > dead, no LEDs lit at all, so this is perhaps just a random failure, rather > than the same pattern. > > Still, they never really admitted that there was a firmware issue causing > the failures. They just emailed me specifically to suggest that I installed > the June update as soon as possible. The notes they sent about the changes > are very vague: > > > - New VLAN debugging tools to help detect VLAN configuration problems > - Distributed layer 3 roaming (no concentrator appliance required) > - 802.11k (radio resource management) and 802.11r (fast BSS > transition) wireless protocols > - Hotspot 2.0/802.11u wireless protocols > - Enhanced stability and performance improvements > > I guess the “enhanced stability” means that MR16s stopped bricking > themselves. I asked specifically about this, and was told that the details > of the update were “proprietary”. Still, it seems to have done the trick. > > Thus, I was surprised to read that the recent firmware updates were giving > you problems, as I’ve not seen this, so far as I know. However, perhaps the > issue is different versions of the MR16 hardware; if everyone else was > seeing the nearly 10% firmware-caused lock-up rate that I saw our first six > months of deployment, I hope I would have heard of it. Thus, I’m guessing > most people didn’t have the “orange power LED” failure; and similarly, > perhaps your issue is only with your specific MR16 revision. (Not, of > course, that the system gives us any info about different possible hardware > revisions, so who really knows. Maybe all MR16s are identical, and it is > just random, or based on some difference of deployment.) > > Anyway, I’m wondering how you are tracking the problem. Do you just get > user complaints when no one can connect, or is there some way to > automatically monitor the dashboard, and reboot APs that have had zero > clients for a long time? (Is there a report that gives daily users per AP? > I have not found anything better than checking the APs list several times a > day, though it would be handy to see peak client counts and such.) Is > there any way to proactively reboot groups of APs early in the morning or > something? > > Thanks for any insights you can share. > > Steve Bohrer > Network Admin, ITS > Bard College at Simon's Rock > 413-528-7645 > > > > > > > ********** 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/.
