WAP321 - device freezing

[leykappp1]

Hello,

we have 2 WAP321 units joined in a cluster and they both keep freezing after about a day of operation (they don't freeze at the same time). When we had a single unit, no freezing occured - the problems have started after second unit was purchased and cluster was formed. The only other configuration change (done at the same time as forming the cluster) was switching from using untagged VLAN1 for both wireless and LAN networks to tagged VLANs, but the cluster seems like more likely culprit.

After the first crash, I have enabled log persistence with level INFO - the log didn't provide any clues though:

(the crash - according to our monitoring - occured at Jun 14 08:05, the device was then manually restarted around 08:18)

Jun 14 2013 08:19:22	info	clusterd[1087]	[shared_config] their(172.16.1.3) key is better than  our(172.16.1.2) key, fetching their data
Jun 13 2013 18:33:43	info	hostapd[1217]	wlan0: IEEE 802.11 STA 00:37:6d:3e:d7:55 deauthed from  BSSID d4:8c:b5:be:6b:10 reason 4: Disassociated due to inactivity

Second device doesn't show any helpful log messages either.

Both units have the most recent firmware version (1.0.3.4).

Diagnostic information from the most recently crashed unit is attached.

Thank you for help,

Pavel

[jeffrrod]

Dear Customer,

Thank you for reaching Cisco Small Business Support Community.

Since the clustered devices attempt to freeze after a day or so of operation, I think this might be addressed by setting  "Automatic" channel assignment which reduces mutual interference (or interference with other WAP devices outside of its cluster) and maximizes Wi-Fi bandwidth to help maintain efficient communication over the wireless network. The automatic channel assignment feature is disabled by default.

Single Point Setup > Channel Management

Please give that a try and let me know if anything comes up and/or if there is any further assistance I may help you with.

Kind regards,

Jeffrey Rodriguez S. .:|:.:|:.
Cisco Customer Support Engineer

*Please rate the Post so other will know when an answer has been found.

[cio]

Hello,

I have same problem. WAP321 freezing (after ~1 day of operation):

1. AP doesnt response to ping;
   
2. WiFi SSID is exist;
   
3. nobody can connect to WiFI network;
   
4. Ethernet interface on switch to which it is connected AP is active.
   

WAP321 info:

1. it has the most recent firmware version (1.0.3.4);
   
2. it doesnt belong to a cluster;
   
3. it configured to use VLANs;
   
4. IP Addres on management VLAN is static;
   
5. PoE doesnt used.
   

AP work properly until next freeze only after restart (power off and on).

According above I think causes of the problem in something else.

Best regards,

Dmitry Gavrilov.

[cio]

Hello!

Yesterday I had this problem with my APs (I have 2pcs. WAP321):

1. First of them freezed twice per one day (every ~4 work hour)!
   
2. Second frezed during night. I dont know when exactly, but yesterday at 6:21PM my iPhone was disassociated due to inactivity and today it failed to connect at 9:00AM.
   

For diagnostic purpose I configured SysLog on second AP and no errors in log (last record is about iPhone disassociation) => therefore AP just freeze!

Additional info about my APs:

1. they were purchased at different times from different suppliers;
   
2. their configuration is equal;
   
3. the model of switches to which it is connected APs is equal (HP V1910-16G) and configuration of Ethernet ports on switch is equal too.
   

Any ideas?

Best regards,

Dmitry Gavrilov.

[Eric Moyers]

Hi, My name is Eric Moyers. I am a Network Support Engineer in the Cisco Small Business Support Center. Thank you for using the Cisco Community Post Forums.

What all do you have configured on them? Captive Portal? Are they Bridged or repeated? Clustered? What firmware are you on?

When they frezze, can you still get to them through a hardwire connection from the LAN? if you can while it is frozen, Log in to the device go to Administrative Tab and Support Information. Download the Diagnostics file and post here please.

Thanks

Eric Moyers    .:|:.:|:.

Cisco Small Business US STAC Advanced Support Engineer

CCNA, CCNA-Wireless

866-606-1866

Mon - Fri 09:00 - 18:00 (UTC - 05:00)

*Please rate the Post so other will know when an answer has been found.

[leykappp1]

Hello,

enabling the Automatic channel assignment didn't help.

When the device freezes, it is frozen completely. It gets into this state gradually - management web interface is significantly slower when getting close to the crash. An hour or two before the crash (freezing), the management interface is no longer accessible (page won't load) and WLAN network connection cannot be established. The device still responds to pings from hardwire LAN network though. Few hours later, the device stops responding even to pings.

Captive Portal is disabled. The clustering doesn't seem to cause the problem, since (in an attempt to pinpoint the error with support engineer) we have tried to disable it without success.

One thing we have noticed is that the device is slowly running out of available memory. We have downloaded Diagnostic information at regular intervals and the amount of free memory is constantly dropping, until you can no longer login (to download diag info).

The only other change (before APs started crashing like this) was switching from untagged to tagged VLANs (this is what we have in common with Dmitry Gavrilov).

We have created CISCO support ticket regarding the issue.

Pavel

[Eric Moyers]

What is the case number for your issue? I would like to look into it.

Thanks

Eric Moyers    .:|:.:|:.

Cisco Small Business US STAC Advanced Support Engineer

CCNA, CCNA-Wireless

866-606-1866

Mon - Fri 09:00 - 18:00 (UTC - 05:00)

[leykappp1]

The case number is 626336969.

Thank you for looking into it,

Pavel

[cio]

Hello, Eric Moyers!

All answers on your question about APs configs you will find in attachments.

I downloaded diagnostic info every ~15 minutes from my APs until it was possible.

Configuration of second of them (it has ID "210" in diagnostic info file name) was reset and it was reconfigured completely.

The start time of the second AP is ~equal the time of first diagnostic info download.

The start time of the first AP (it has ID "211" in diagnostic info file name) for one hour before second AP.

As I wrote above, after WAP321 freeze I cannot connect through WiFi or Ethernet.

But I can see SSID from any other WiFi devices and Ethernet is up from switch point of view.

Before WAP321 completely freeze I was able to connect to the admin WEB-interface, but was unable to download the file with diagnostic info.

Then I was able to ping APs but unable to connect to the admin WEB-interface. And finally it freeze completely.

Time before freeze (for bouth APs) is ~3-4 hours!

As you can see in diagnostic info files the size of free memory was reducing very fast!

I think the reason is a memory leak in some process in configuration with VLANs.

Also, I opened trouble ticket (Support Case Number 626416237).

Any ideas?

---

Best regards,

Dmitry Gavrilov.

PS. BTW, I found discussion about similar problem here https://supportforums.cisco.com/thread/2199385

Tomorrow I will try to check this solution.

[cio]

Hello,

I cheked solution and I beleive it is works: my APs uptime is more then 4 hours and they have almost constant level of free memory.

---

Best regars,

Dmitry Gavrilov.

[coredialeng]

Dmitry or anyone,

What was the solution per you ticket? We've purchased 4 of these devices already. After setting things up everything works fine. After a while there is not ping response. And just like you, people couldn't connect to WIFI. In all cases, a reboot magically resolved the issue.

These are by far the worst APs I've ever worked with. The Cisco Aironet 1600s are well worth the money and saved hassle.

[mpyhala]

Hi CoreDial,

Make sure the Untagged VLAN matches the PVID or native VLAN of the port on the switch. There appears to be a native VLAN mismatch in these cases that is causing the issues.

- Marty

[cio]

Hi CoreDial,

And make sure that you does not disable Untagged VLAN option!

---

Best regards.

Dmitry Gavrilov.

PS. Until now uptime of my WAP321 is ~9 days => solution works!

[coredialeng]

Hi,

I disabled the Untagged VLAN option and set the management VLAN explicitly. I'm not seeing how this can be a native vlan mismatch in this scenario. If the AP is tagging traffic to the managment IP with VLAN X, the switch just need to have pass traffic for VLAN X (e.g. switchport access VLAN X, or switchport trunk allowev vlan add X ).

Also, if it works initially, why does it randomly stop working. It seems to be a bug/memory leak as described in earlier posts...