Meraki

thomasthomsen

Old UI on switches seems to "work"... But other stuff times out, like the "staged switch upgrade wizard". Loading the "list" of switches or APs takes forever. Loading list of inventory takes forever. And so on....

thomasthomsen

Since yesterday (we think) this shard has been acting up, at least for the 4 customers I have so far located on this it. Case created, confirmed that something is going on (from other cases created), they are working on it.

thomasthomsen

9176 seems to do the "right thing" in Meraki mode, when I look at my local ones. I get 4SS on 5 and 6Ghz, and the expected 2 on 2,4Ghz with the PoE+ PoE i deliver. And I dont think I have seen the Cisco classic bug where it only runs 1Gbit on PoE+, my local ones run 2,5Gbit (as expected) on PoE+ (Currently running latest MR Beta on these).

thomasthomsen

Well yes, I have had "many" CW9172I APs at different customers , with different switches, where we fx. unplugged an old MR36 or MR44 AP, connected a CW9172I , and it never powered on. I have a customer with MS130-8P-I where we have seen this plenty of times. The other customer where we have this problem might be because of "the MS120 Q2xx PoE problem". The 3rd customer where we have encountered the problem had MS150 switches, and there was also something with PoE on those switches. But in a nutshell, the CW9172I never seems to power on, if you leave it there, it might power on hours later, seemingly randomly. This could also happen if you rebooted your switch. After upgrading customers to the latest (or before the latest, Im not 100% here) 32.1.6 MR software the problem seems to have gone away.

thomasthomsen

Well, you tell me ? 🙂 Because when I moved those APs to my home lab, with a different switch, they got online. When I then moved them back to the customer site (most likely after a software update) they worked as expected. What we did see was that this batch of CW9176I started to boot on the C9300L-24P, seemingly getting a connection to the dashboard (from packet captures, not from the LED status), and then, for some reason, they rebooted (lost power), seemingly without any reason. (there was no indication on the dashboard that there was a connection, but we could see the AP change its hostname on LLDP, so something happened). And then they started over, and over, and over, and ..... Of course I cant tell you 100% if it was a PoE problem, or something else, but clearly (at least to me) something was "wrong" in the switch / AP combo , until the APs got online on another switch, and got a software upgrade. I might just have "PTSD" from all the CW9172 PoE problems I encountered, and where there seemed to be no real answer - I mean, I now have an answer for the MS120 switches, and I think i have an answer for the MS150, but not really for other switches I encountered CW9172 problems on.

thomasthomsen

Actually come to think of it. I do have some more information. we have been installing the CW9176I at the same customer, same switch, same switch software. And the first 15 sites no problems, but then it started. So Im betting that we had open a new box (they come in boxes of 5) of CW9176 that might have been a little bit different. - Just a thought. The first AP where we had the same problem, we tried a lot of things, and after 30 - 40 mins, it actually just came online, without us having done any changes to configuration or just troubleshooting, like unplugging the AP, packetcaptures, cable tests, cycling the port and so on. - But the second AP from that "carton" we never got online, and we also tried the 3rd AP, also never got online (this was another site, this customer has many small offices, so almost only one AP pr site, but same model of switch and software). - We created a case, and support told us RMA, as mentioned, but I was curious, so I took those two APs back to the lab got it online, and now it works fine on the customers site.

thomasthomsen

Its C9300L-24P-4G running 17.18.2 in "meraki mode" (I cant get used to the "new terminology".) , and the APs are running ? ?? - I dont know, because they are new from the factory. If they ever got around to downloading new software, it would be the latest recommended 32.1.6 , but the AP never gets that far. When I got them upgraded, by having them connected to a MS130-12X running the latest MS18 software, and brought them back to the C9300L-24P-4G they booted and connected, and have stayed online as expected.

thomasthomsen

Some of it was being discussed on the Cisco Community side at some point. https://community.cisco.com/t5/wireless/cw9172i-increased-cabling-sensitivity-poe-detect-issues/td-p/5329049#:~:text=CW9172I%20has%20increased%20cabling%20sensitivity%20and%20experiences,or%20%22power%20given%20but%20not%20good%22%20errors. I have just had the experience that the CW9172 failed not powering on on MS130 and MS150 switches (non mGig versions). - But all of a sudden , after hours, it would just boot. We have seen, on a previous release, cant tell you which one, that when we upgraded the CW9172, and they rebooted, they would be stuck in the same manner, not getting PoE. It seems, that CW9172 on the latest recommended release are ok. But as mentioned, if you get a new shipment, that might contain older software, you might be in trouble. So I think the CW9172 problem is "resolved", but Im worried about the CW9176. This is not the same problem of course, because the AP powers up, but at some point, just shuts down, and starts over.

thomasthomsen

I have seen the problems with CW9172 not powering on at all on multiple different switches, not just MS120 (and the MS120 is its own thing). - Also on MS130 and MS150 switches. The fun thing is if you leave the AP on the switch (where it still does not get power), then perhaps after a couple of hours, it will power on, and connect. (I think I found a bug about this yesterday It seems ... seems, to be fixed on the latest recommended release, but again ... if you get a new AP from the factory, what software does that run ? The 9176 I have only had a couple of singular times, but always, it seems, that it actually get power, but then, after a short while, it just reboots, and then repeat. Mostly (always?) on C9300 switches - I have not that many customers with CW9176 and MS switches.

thomasthomsen

I have also never had any problems with CW9178.

thomasthomsen

This could of course be it. - Because I had the 9176 that supported demed RMA ready, connected to my home lab MS130 (where they booted, and where happy), and now today connected them back to on the same site that had problems yesterday (with more then one AP), and now it works on this site. The site have C9300L-24P-4G. The problem here is of course that you do not know what you get (Like a box of chocolates ? 🙂 ) from a new shipment of APs 🙂 - So you can potentially be in some serious problems if you dont have another switch to "pre-deploy" them on.

thomasthomsen

So ... after now having deployed some / a few / lets say a couple of hundred ... CW9176 and CW9172 APs there seems to be "something(tm)" going on with PoE for these devices. But no-one seems to talk about it ? - Or am I just the unlucky one ? 🙂 I have learned that there is a specific issue with CW9172 and certain MS120 switches (I dont know if Im allowed to discus publicly, there are known BugIDs on this - go look if you have access). I have also fallen into the trap of thinking it was the AP (CW9172) because of PoE issues with the MS150 (resolved through software it seems, not the mentioned -MP PoE++ problems from the release notes, but "something else" with PoE on that switch model). But what is going on with the CW9176I's that I have deployed. Usually they just power on fine, connects, works, no worries and everyone is happy. But once in a while, they seem to power on (led on) and starts to go through their boot and discovery process, but then, just as the first few packets are exchanged with the dashboard (we can see this on packet captures) , it reboots, seemingly loosing power, and starts over. Moving the AP directly to the switch (aka. to make sure its not cabeling and even switchport) does not seem to resolve the problem. From cases created, there is nothing to see on the "Meraki end of things", and RMA was suggested from support. This was happening "often" / "only" (cittaion needed), on Catalyst 9300 switches. "For fun" i took the failing APs to my lab, connected them to another switch (MS130-12X) and they booted and connected just fine. Have anyone else seen this ? Can I have my sanity checked here ? 🙂 (And of course I should have done some PoE debugging, I know this now in retrospect, but still .... 🙂 ) Thanks.

thomasthomsen

Have anyone seen the C8121 in CCW yet - I would love to see the price of that one. C8111 seems to be there.

thomasthomsen

Like everyone else (in this tread 🙂 ), we are just waiting for someone who "dares" to be the "first mover" 🙂 We also have customers with C9500 running in the "old monitor mode", and we also want to go Full Cloud Configured WITH StackWise Virtual .... but .... who dares 🙂

thomasthomsen

Lets see if they will follow the current Cisco classic standard for license tiers. Im guessing ... yes.

thomasthomsen

Isent the new Meraki SUB licenses C8K-G2-MX Small - LIC-CSR-S-E/A, and C8K-G2-MX X-Large - LIC-CSR-XL-E/A just essentials and advantage like on the current MX small , medium , large ect models in SUB ? I mean, sure, compared with the "box" running IOS, there are more license differences, but within "meraki" I dont think there will be any difference (except the new ones here do not support MI).

thomasthomsen

Do we know if these models support dot1x on the LAN ports ?

thomasthomsen

I was in a "panic" because the neighbour advertised the default route. But you are right, the networks needs to be specific. But it still installed the default route when I had a 10.100.0.0/16 filter 😕 - With all the benefits of it being a higher priority for all spokes in autovpn ... thats where the panic came from 🙂 So something is definitely going on there.

thomasthomsen

RIP MX6x and MX75 when I look at price, license price, and performance. But that was expected I think... I mean, anyone who saw these small C8K routers must have thought ... thats a new small MX 🙂 The two C81X1 belongs to the C8K-G2-MX Small SUB SKU i see. C8455 to the C8K-G2-MX X-Large SUB SKU ..... So .. can we now easily guess what SKUs will be in the , surely, upcomming Medium and Large SUB SKUs .. I think we can 🙂

thomasthomsen

---- deleted previous message ---- I was wrong about what I wrote. The filter needs to be specific. When entering fx. 10.100.0.0/16 I dont get, lets say 10.100.24.0/24 as a route. So it needs to be specific. The "second" i put in 10.100.24.0/24 its received and installed. And if your neighbour advertises a default 0.0.0.0/0 route, its installed, regardless of the filter 😕

thomasthomsen

I see all the mDNS packets from the wireless device, on the wired side of the AP 🙂 But apparently they are not forwarded on the radio, almost like P2P blocking on the wireless side for mDNS. I made a case, they pointed me to a known mDNS issue on the latest MR32.2.1 When I downgraded everything worked as intended. (It was not just airplay, also other mirroring / casting devices - but all related to mDNS)

thomasthomsen · ‎Jan 27 2026

I figured it out, there is a bug in the latest MR32.2.1 that breaks mDNS. Back to MR31.1.8 and everything works as expected.

thomasthomsen · ‎Jan 27 2026

For fun I have now played around with the "Microsoft Entra ID Integration with Splash Page" https://documentation.meraki.com/Wireless/Design_and_Configure/Configuration_Guides/Encryption_and_Authentication/Microsoft_Entra_ID_Integration_with_Splash_page - I know, not really Access-manager related, but still, people might want to know. You can authenticate using this method, using Entra External ID, but ONLY if you use the UID "email" as the username it seems .... And those are usually ..... "smashhandontokeyboard-smashhandontokeyboard-smashhandontokeyboard-smashhandontokeyboard@external.onmicrosoft.com" PS: The UID does not work with Access-Manager and EAP-TTLS ( I already tested 🙂 ).

thomasthomsen · ‎Jan 27 2026

Thanks, that was super helpful, and just the info I needed 🙂 Do you know if it will "allow" a smaller subnet to be recieved then the actual subnet permitted ? So, for example, I permit 10.10.0.0/16, but the router on the other end advertises 10.10.0.0/24, would 10.10.0.0/24 be accepted ?

thomasthomsen · ‎Jan 27 2026

Do anyone , by chance, know of some documentation on the inbound route filter part ? What is the correct input in that field ? A network like 10.0.0.0/8 or ? (I have not tried it yet 🙂 ) /Thomas

About thomasthomsen

thomasthomsen

Thomas Obbekær Thomsen

Community Record

Badges