Random Rebooting of Cisco 1252 Wireless Access Points

[stephenshaw]

Hi,

I'm trying to see if anyone else has run into this problem, and, if so, suggestions to try in order to resolve it.

• We have 50 - 60 cisco 1252 standalone wireless access points - all of which do random reboots
  
• All of the 1252s use the AIR-PWRINJ4 power injector since the 3560 PoE switches do not have sufficient power for two radios
  
• All of the injectors and the switches are connected to UPS systems
  
• Switch logs only show the typical up/down interfaces status
  
• I had read on this forum that with the older AIR-PWRINJ3 that is was recommend to configure "power inline never" since the PoE switchport was interferring with the injector. I have now configured this on several ports. This did not solve the problem
  
• Original IOS code used was 12.4(10b)JDA3 and about 10 of them were upgraded to the latest and greatest (at the time) to 12.4(25d)JA1. This did not solve the problem
  

The reboots could occur within 24 hours or go several weeks or even months before a reboot happens again. I do have Cisco investigating but, to date, they have not seen this before. I'm suspecting it's a power injector issue rather than with the actual access points.

Has anyone run into this with the standalone 1252s using power injectors?

Thanks,

Steve

[Nicolas Darchis]

The console output when the AP is rebooting would be the best ressource to understand what's happening.

Otherwise the "show log" of the AP is also an interesting thing to check, rather than the logs on the switch.

[stephenshaw]

Thanks Nicolas,

we have no syslog and, as such, when the access point reboots, any info prior to the reboot is lost. And, when I indicate the reboots are random, that's exactly what I mean - how could we possibly know which one is going to reboot in order to view the console output?

However, you bring up a point about the log. We'll see about trying to setup a syslog server to hopefully gain some useful information.

Cheers,

Steve

P.S. I have now setup 5 access points to send the logs to a temporary syslog server. Now, I want one of these 5 to reboot. ;-)

[Leo Laohoo]

All of the 1252s use the AIR-PWRINJ4 power injector since the 3560 PoE switches do not have sufficient power for two radios

The latest IOS will allow the 1250 to operate both radios with 15.0w of PoE power.  The "catch" is that the last two MCS rates will be disabled.

we have no syslog and, as such, when the access point reboots, any info prior to the reboot is lost. And, when I indicate the reboots are random,

Fortunately, it does.  IT will tell whether the WAPs rebooted due to a software or hardware crash or a power failure.

If power failure, I'd recommend that you plug the WAP into the switch directly and observe whether or not you still see issues like this.  Just to clarify that the exact model of your 3560 is a 3560G or not?

Again, if it's a power failure can you visually inspect the power plug connected to the injector?  There are some manufacturing batches where you need to push the power plug HARD into the socket of the injector.

Another thing about power failure, if you have a faulty patch cable to the WAP, vibrations can cause the WAP to loose power.

[Kayle Miller]

How old are the 1252's?  are they connected with Cat6 patch cables?

The reason I ask and it's a longshot but I had a client that was using 1252's with cat6 patch cables and they had random reboots aswell. When we pulled the AP's out of the ceiling we weren't able to replicate the issue until we realized that we were using cat5 cables on the bench. We pulled on of the cables from the ceiling we were then able to replicate the issue. The slightest movement or virbation would cause the AP to reboot. we replaced the cables and the client never had any other issues; and its been close to a year.

the only thing we could figure was that there was something different with the RJ45 connector that casued it to not sit tight and stay in place in the jack on the AP.

It's worth looking into.

[stephenshaw]

Hi Kayle,

interesting, I had already checked the connections at our head office - no issues found. None of our sites use Cat6 cabling ... it's all Cat5. The 1252s vary in age from a few years old to only 4 months old.

Thanks,

Steve

[Kayle Miller]

Stephen,

     It might be worth taking one of the AP's and just wiggling the cable and see if the unit reboots, if not then it has to be either software or environmental like heat or corruption.

[Leo Laohoo]

Can you please elaborate what exact model is your 3560?  Is it a plain 3560 (including v2) or a 3560G?

If the answer is a 3560G, then you can verify if you have a layer 1 (cabling) issue by performing a simple TDR test.

Here's the command combination:

1.  Command:  test cable tdr interface ;

2.  Wait for about 5 to 7 seconds;

3.  Command:  show cable tdr interface ; and

4.  Please post the output.

[stephenshaw]

Hi Leolahoo,

we're using V2 WS-C3560-48PS and 24PS.

Physical connections have been checked with no problems found (and that includes unplugging from the power injector and plugging back in). All of the 1252s and only the 1252s experience random reboots. The 1252s are used throughout mulitple offices across North America, all using Cat 5 cabling.

No "bugs" have been found but we upgraded IOS code (which did not resolve the problems).

We have Cisco invetigating and should hear back in the New Year.

Regards,

Steve

[Leo Laohoo]

we're using V2 WS-C3560-48PS and 24PS.

Bad luck for me.  I was hoping it was a "G" model so you can run a TDR.

Can I make a suggestion?  May I request if you can take one of WAPs that keep rebooting and relocate it to yoiur desk for testing.  What I'd like know is what happens when you leave a console cable and terminal running.  I want to capture the logs as it happens to determine as to why your WAPs keep bouncing.

Another thing, you said the physical connections have been checked.  How was this checked?   The reason why I'm asking (and the reason for the TDR) is because alot of cheap so-called "cable testers" will only test the first two pairs of an ethernet cable.

You said PWR-INJ3.  That's another possibility.  We have around 200+ of these.  And during initial installation, we discovered that the power plug, if connected without exerting any effort, will come loose.  Can you check to ensure the power cable connected to the injectors are inserted tight?

But it still does NOT make any sense as to why ALL the 1252 are rebooting and ONLY the 1252 are rebooting.  What makes it confounding is that they affect multiple sites.

IF you can get a "sh version" of a few of them that have rebooted we can determine if the WAPs rebooted due to an IOS bug, hardware failure or simple a power reset, this will be a good first step to the troubleshooting.

PS:  I don't know if it would help if you can post this thread to the Cisco TAC so both you and Cisco TAC won't be repeating the story and troubleshooting un-necessarily.

[Leo Laohoo]

But it still does NOT make any sense as to why ALL the 1252 are rebooting and ONLY the 1252 are rebooting.

It actually does.  And I'm an id10t for not seeing it.  Let me explain:

All of the 1252s use the AIR-PWRINJ4 power injector since the 3560 PoE switches do not have sufficient power for two radios

So am I safe to presume that you have other model of WAPs that can operate at 15.4 W PoE?

So why the 1252 WAP and not the other model?  Because all the 1252 WAP having this issue are all powered up using a power injector.

This brings me to my initial post that the issue could be the power cable and the injector.  I've personally seen my past implementation of the power injectors.  One could mistake that the power cable is inserted correctly until you apply a little bit more presure to push the power cable into the injector socket.  You can get about 4 mm more.

How about a test if my theory is correct?  If your WAP is running the latest IOS then it's safe to power up the WAP using 15.4 w PoE.  Plug the WAP directly to the switch and observe if the 1252 still reboots randomly.

Let me know how it goes.