MR26 AP's Unresponsive (Even After Replacements)

[aluminumfoil]

I'm doing contract work for someone with a large number of MR26 AP's. I was called out when two of them went down and it appeared the PoE feature was no longer functioning so I had the cabling team come out to test for faults - but they found nothing out of the norm. I returned on site and tried powering on the AP's via an AC adapter and got them up and running - connected to the switch and returned to normal function. I then tried plugging them into different PoE cables (without AC power) and just heard a series of clicks and buzzes but no LED action...I reported it to the Meraki team and they said the units had likely gone bad and they sent out replacements. They were received while I was out of town and hooked up, despite my objection. I took a quick peek at the portal the same day and it showed the devices as powered on but without internet connectivity. I looked at the logs at that time and noted a couple of devices successfully connecting to them for the first few minutes after they were powered on. They have "died" again since that time and I'm left scratching my head. If the AP's are new and the cables are in good shape then what's the problem? GAH! I can't understand why the portal can't just save the logs from the AP's should they go down -- how am I supposed to work backwards to troubleshoot issues if I can't examine their logs?

[PhilipDAth]

When they are not working what does the local status page report?

https://documentation.meraki.com/zGeneral_Administration/Tools_and_Troubleshooting/Using_the_Cisco_M...

Meraki are using more data centres now then they used to.  Is there any firewall limiting what address ranges and ports they can talk to?  If so in the dashboard go "Help/Firewall Info" to get the latest ranges.

[aluminumfoil]

---

@PhilipDAth wrote:

When they are not working what does the local status page report?

https://documentation.meraki.com/zGeneral_Administration/Tools_and_Troubleshooting/Using_the_Cisco_M...

Meraki are using more data centres now then they used to.  Is there any firewall limiting what address ranges and ports they can talk to?  If so in the dashboard go "Help/Firewall Info" to get the latest ranges.

---

Thanks. Didn't know about that. Clicked the linked but it's only showing the results for the AP I have here at my office (I also have admin rights to theirs under my account). Is there a way to switch over to that account or is this something that needs to be accessed on site?

[PhilipDAth]

It has to be accessed on site.

[PhilipDAth]

Also what is actually supplying the PoE power?  Perhaps the device supplying the power is the issue.

[aluminumfoil]

---

@PhilipDAth wrote:

Also what is actually supplying the PoE power?  Perhaps the device supplying the power is the issue.

---

That was my initial thought but between having the cables tested and hooking up a laptop to them and having PoE detection kick in and gaining full network access I'm not sure the cabling is the issue. It's also a head-scratcher as to why the new devices worked temporarily. If the voltage decided to go haywire could the devices' PoE functionality get fried?

[PhilipDAth]

Switches have a PoE budget, for example 370W.  If you have devices turning on or off during the day, or using more power then that reduces the PoE budget for other devices - such as access points.

Here is an acid test.  Plug the access point into a power transformer.  If it then works perfectly then you have to consider an issue with the PoE supply.  If it still fails with a power transformer then something must be wrong with the access points.

[aluminumfoil]

---

@PhilipDAth wrote:

Switches have a PoE budget, for example 370W.  If you have devices turning on or off during the day, or using more power then that reduces the PoE budget for other devices - such as access points.

Here is an acid test.  Plug the access point into a power transformer.  If it then works perfectly then you have to consider an issue with the PoE supply.  If it still fails with a power transformer then something must be wrong with the access points.

---

Thanks again... 🙂 I don't imagine any issues powering them up with AC again - though I'll check. I don't know a great deal about PoE limits so that sounds like a promising explanation - though it posses the question of why only two specific AP's are taking the brunt of it if that's the case. Guess I'll be calling the cabling team back for another look.

[PhilipDAth]

Try swapping where the problem switches plug into the PoE switch with some APs that don't have an issue.  See if the problem follows the ports being used, or stays with the APs.

If you google the model of switch being used you should be able to find out what the PoE budget is for them.

[aluminumfoil]

---

@PhilipDAth wrote:

Try swapping where the problem switches plug into the PoE switch with some APs that don't have an issue.  See if the problem follows the ports being used, or stays with the APs.

If you google the model of switch being used you should be able to find out what the PoE budget is for them.

---

Thought about it. Just concerned they could be fried too. If I give it a shot I'll hold you responsible for it. 😛

[PhilipDAth]

It might be easier to just buy a nice new Cisco Meraki switch ...

[aluminumfoil]

---

@PhilipDAth wrote:

It might be easier to just buy a nice new Cisco Meraki switch ...

---

That's really not a bad idea. I imagine they would work impeccably well together. I'll look into what options exist and their pricing -- might be worth it over my fees and those of the cabling team having to do repeat visits. I'll be back to bug you if I need someone to run them by. 😉

[Adam]

We had a similar issue to this and went round and round.  Replaced the AP.  Moved it to other ports on the switch.  Had the cable tested and was told repeatedly that it was good.  I even bought a cable tester and tested the cable myself and the pinouts were good.  But inevitably the AP would go down.  Intermittently.   Did and RMA and the new AP did the same thing.  Finally I plugged the AP directly to the same port on the switch using a small patch cable in the rack and was able verify if it was the run going to the AP.  I had the installer pull a new cable and the AP has been rock solid ever since.  May not be your issue but POE stuff can be finicky.  Could be a small nick in the cable or running over florescent lights.  Lots of factors that are all worth considering. 

[PhilipDAth]

When testing the cable you really want something that can do a TDR and impedance test.  Only the more expensive test tools can do this.

ps. Meraki switches can also do this kind of test.

[aluminumfoil]

---

@Adam wrote:

We had a similar issue to this and went round and round.  Replaced the AP.  Moved it to other ports on the switch.  Had the cable tested and was told repeatedly that it was good.  I even bought a cable tester and tested the cable myself and the pinouts were good.  But inevitably the AP would go down.  Intermittently.   Did and RMA and the new AP did the same thing.  Finally I plugged the AP directly to the same port on the switch using a small patch cable in the rack and was able verify if it was the run going to the AP.  I had the installer pull a new cable and the AP has been rock solid ever since.  May not be your issue but POE stuff can be finicky.  Could be a small nick in the cable or running over florescent lights.  Lots of factors that are all worth considering. 

---

Rockstar response. I'll relay your post in full to the cabling team. Thanks for sharing!

[aluminumfoil · Accepted solution]

Turns out the original AP's were bad after all. The snag was the phone rep implying that the new ones would be plug-and-play as far as keeping settings for the replacement devices. Instead of simply plugging in the new devices, they needed to be registered in the portal and set up again. Thanks again for all the responses and help!