Losing access to network

[JesseR]

Experiencing similar issues on an MSM760 with (25) or so MSM 422s.  Devices are typically all Macbooks or iPads.  I've seen when we get a lot of active connections on an AP (for example 40-65 client connections, half on the 2.4 radio and half on the 5 radio), that internet access becomes intermittent.  Client SNR's are still strong (20-35 SNR value)  and we never really see disconnects when looking at the logs/web interface on the MSM/AP itself.   Yet these seem to be losing their network connection...

Controller code is 5.5.3.0 at this time and considering upgrading to 5.7 (Procurve tech emailed me the code) but since the release notes aren't available I'm hesistant to do this....

All my 422 APs are set with a static channel on the 2.4 radio, (either 1, 6 or 11).  We do not use wireless security filters in the VSCs.  I'm seeing this on non-access controlled VSCs (can't verify whether it happens on the access-controlled VSC).

Can anyone verify a fix in 5.7 that's actually worked for them regarding loss of access?

[ISoliman]

Release note is in the below links:

http://bizsupport2.austin.hp.com/bc/docs/support/SupportManual/c03127429/c03127429.pdf

[JesseR]

So one of the things listed in the 5.7 Release Notes under "ISSUES" is:

"(Applies to MSM310, MSM320, MSM325, MSM335, MSM422.) Sometimes an Apple Mac

Book is unable to forward traffic to the external network even though the wireless signal is shown

as Excellent. Furthermore, the wireless driver may stop working. Either restart the AP or

re-synchronize it after a radio configuration change."

I'm guessing that means the issue is still an outstanding problem since its listed under ISSUES and not under FIXES?

I dont' see anything in the release notes that makes me think the problems experienced here in this thread were addressed in 5.7?  Unless I missed a specific line item entry in the Release Notes? 

[Scott Farrand]

This seems to have been a problem for a while... (not sure if it was manifesting in 5.4 code, but it's seemed to have been an issue in all of the sub releases of 5.5.x code I've tried)...

At this point, it'd be very nice to have an option to reboot all ap's at the same time...

[rsaeks]

From what it sounds like - the issue seems to be when the radios change configuration a restart is needed.  Does that mean if they are set to auto channel or power with an interval there will be an issue until that AP is restarted?

[temuri426]

hello, 

my problem appeared again..  :(

clients are receiving ip addresses from dhcp server, but no l2, l3 conectivity..   In controllers log files are these errors : can't identify discovery port, this apppears several time, after rebooting WLS's everything works good. 

Is it possible that it maybe AP's problem ? because If controllers are in down state, clients must be capable to conect and browse without WLSs? 

[Kell van Daal]

I would like to urge everyone that is experiencing this problem to open a case with HP support.

To quote a colleague who said it better than I can:

" The posts in this thread describe an issue which is clearly neither expected nor desireable behavior. If you are experiencing this issue, please log a case with HP Networking support. Sometimes, logging a case provides key data points or other information that speeds time to resolution. Other times, it might allow you to participate in verification of the fix. At the very least, logging a support case will ensure that you are notified when the fix is available (sometimes even before the fix is publicly posted).

No one likes to experience an issue that creates so much frustration - but calling support and partnering with us toward a resolution is win-win.

Best regards,

Lynn-Marie

HPN Global Support Engineering"

Also it would be helpful to know what the support case numbers are (even if the case has been opened earlier), so we can correlate the issues and compare. There might be different issues going on. If you are not comfortable posting the support case number in a public forum, feel free to send them in a private message to me, or email them to me at Kell.van.Daal@hp.com referencing this thread and your username.

Rest assured, we want this issue fixed as much as you do, but we need as much information as possible to make that happen. Opening a case with HP Support is the best way to get that done.

To log a case online, go to: https://h10145.www1.hp.com/help/help_questions.aspx?l2id=48&SelectedTab=3

To log a case over the phone, go to the following site to find the phone number for your country: https://h10145.www1.hp.com/help/Help_ContactInfo.aspx?cwp=2&SelectedTab=2

[rsaeks]

I emailed HP support about this and have a case number: 7637982326.

[alansev]

My school are intending to buy an HP wireless solution. This thread raises issues for me. Could I ask if contributors think it's been resolved and if you consider HP wireless support to be good in general?

[ISoliman]

Anybody getting errors similar to the below:

Feb 21 09:50:47 err  rpcd         CNXXXXXXX BRIDGE_add: BRCTL_ADD_BRIDGE returned -1

Feb 21 09:50:48 err  confighandl  CNXXXXXXXX Network IF ioctl error r1v1.1-4094, cfgerr:(100) syserr:(Network is down)

Feb 21 09:50:48 err  confighandl  CNXXXXXXXXX Network IF ioctl error r2v1.1-4094, cfgerr:(100) syserr:(Network is down)

Feb 21 09:50:49 err  eapolserver  CNXXXXXXXX Interface r1v2 doesn't exist

Feb 21 09:50:50 err  eapolserver  CNXXXXXXXX Interface r1v3 doesn't exist

Feb 21 09:50:59 err  eapolserver  Last message repeated 1 times

Feb 21 09:50:59 err  confighandl  CNXXXXXXXX Network IF ioctl error r1v3.1-4094, cfgerr:(100) syserr:(Network is down)

Feb 21 09:50:59 err  confighandl  CNXXXXXXXX Network IF ioctl error r2v3.1-4094, cfgerr:(100) syserr:(Network is down)

CNXXXXXXXX is the AP serial number which I have covered, to check these kind of logs you have to go to "Controlled AP" then got to "Tools" and from there you will be on the logs window, check the logs if facing the same logs then this might be the cause.

Please all check and let me know, make sure to check the logs after the issue of clients loosing connectivity happens.

[JesseR]

I am seeing those error messages in my Controlled AP logs too, however... I only see those errors for the (1) MSM430 AP I have in place, and the errors occurred just after I rebooted the MSM controller last night. The errors haven't shown up since last night however, and no errors like those showed for any of the MSM 422s.

This is with code 5.5.3 on the MSM760. 

[nogirs]

As an update to my previous post last month, I'd like to report that since upgrading to 5.5.3 (should I really say this?) - knock wood - that our system has stabilized significantly. We have two non-teamed 765zl controllers and about 120 ap's on each. As a university, we have a vast mixture of clients and devices - all seem to working rather well for now. I wish I could say that I did some magical change to the configurations or something to help the others who are losing network, but I can't.  Somehow, the frequent system reboots and losing network connection issues have seemed to have subsided. 

I am not seeing any of the error codes in the configuration logs that the others are reporting.

[ISoliman]

To be honset for me after upgrading to 5.5.3 the network is better than before except some clients disconnect randomly which is giving me a bad time solving it, no solution till now for this

[petrus_kl]

We're experiencing the same issues with 5.5.3. I'm hoping to install 5.7 soon.
We have 80 MSM430 APs and two MSM760 in a team.
My logs are only filled the "Unable to update interim traffic", not the errors ISoliman posted.

Is there a way to restart all APs from the GUI?

[ISoliman]

To restart or even reset the AP from the controller go to "Controller AP" then click on the AP you want to restart, then go to "Maintenance" from there you will find options to reset or restart :)

Anybody check if the "power save mode" in the WLAN somehow is affecting this ???

[ISoliman]

small thing to check also please, does this connection happen (or the ! mark showing) when the client is somehow IDLE ??

For those who are facing this ! mark showing, try to open Internet Explorer and see if that restores the connection or not, it worked with us, ping doesn't restore the connection only reassociating or openning IE Window.

[rsaeks]

Any issues I've been having with clients experiencing conectivity issues are always on Radio 1 for me that is running 802.11n at 5GHz.  If I turn off N, the client connects in the 2.4GHz band and works just fine.

The odd thing is a client will obtain a valid IP address, but the traffic just isn't going anywhere and hte device is not pingable on the network.  This is in 5.7.0.2.

[blackstone]

We had the same problem, loosing network connections on most of the computers.
I’m saying on most of the computers because the older computer with only G standard worked without any problems.
The solution for my company was to disable the 5 gigahertz frequency, not the N standard because N can also be used in the 2.4GHz band.

To do so disable "Radio 1" as a Wireless Service.

I know that the 5GHz is better the 2.4GHz band "As long you are not working beside an airfield". But I rather have a slower working network then a faster that doesn’t work at all. =)

We are having an MSM760 Controller with software version 5.7.0.2.

[amgorb]

We have the same problem here. 6 MSM410 and 2 MSM460, many different Mac users (macbooks,ipads, even some imacs), users losing connection randomly, disabling 5ghz and n-mode doesn't help. Updated firmware to 5.3, 5.5, 5.7, didnt help either. No messages in logs, no way to track down an error or to reproduce it.

Greatly disappointed with hardware choice for wifi network setup.

[AmanaLiving]

Add another +1 on this. 5.7.0.2-01-10750 Firmware.

Case number: 4638703129

[AmanaLiving]

I have resolved the issue on my end.

I was directed by HP support to perform a wireshark capture of the AP, when looking at the wireshark trace, I would notice that the first 3 or so clients would make a DHCP Request, and they would receive back a DHCP Offer.

Subsequent clients would connect to the AP and send out the DHCP Request but not get back an offer from out DHCP servers. After some time, the original client would stop transmitting and another client would then connect get an IP address and the previous client was blocked.

It turns out there was port security configured on our Procurve 5412zl. The default configuration that was on this switch, it was set to Static, with an address limit of 3, and an action of block and send trap.

So it would only allow 3 MAC addresses then it would cut off access on that port.

I had to change port security from static to Continuous to allow all the clients to connect. Now all clients have an IP address and are working normally.

Hope this helps someone think about this issue from a different angle.

[cenk sasmaztin]

Beamforming ability can't resolve this issue this issue continuing 5.7.0 software

[Craig-Lyndes]

Finally got all the support agreements in place (horrible experience!) and had a high profile training over the weekend fail because the access point stopped forwarding packets between the wireless and the ethernet.

Opened Case # 4639380148

Approaching 24 hours with no response.

Case description...

Have had a problem since we purchased this system with the access points not forwarding packets from the wireless network to the ethernet. Have a MSM760 controller and 32 msm410 access points and 3 msm430 access points covering a 50 classroom school with very high density (1:1 - potentially 60 devices per room) wireless. The symptoms are that people connect to an access point, they see Excellent Signal Strength, they see the correct SSID - on the controller side I see the access point, I see a number of devices connected, no errors in the logs - but the users can not communicate with anything on the ethernet. I can restart the access point from the controller, which solves the problem. I can power cycle the access point, which solves the problem. When this first appeared I stopped it by setting the msm410 access points to G only. This was successful until last fall when I added some msm430's to rooms that regularly had a high density of clients, the msm430's can not be set to G only. I needed the bandwidth supplied by N, so now all the msm410's are set to G and N. Since software 5.5.3 the problem has gotten much better, but it still occurs. One factor in triggering it seems to be the presence of a Mac wireless device. Another is high utilization, lots of clients. When both of these are present I am still getting access points that turn into data black holes.

I was told that a high profile training that took place here this Saturday was a failure because no one could connect to the network wirelessly. I took my Win XP notebook to the location and connected to the access point, but could not connect to any web sites. I power cycled the access point and everything worked fine. The woman whose desk is right under the access point said she had not had any problem today, but she is using a Mac. The room across the hall from the location that could not use wireless is served by a different access point and people were able to use it throughout the weekend with no problem.

[sjordet]

Craig-Lyndes seems to have the excact same issue as I do. And rsaeks have made the same observations as I did - with g-only there is really no problem. But we have 130 AP's, where almost 100 of them are MSM410, 7 MSM422, 3 MSM322 and 20 MSM430...

We could really need the extra bandwidth provided by n. Rebooting the AP always helps. Running 5.5.4.0-01-10665 on an MSM750.

[gglater]

There are a lot of messages here about a lot of different products over quite a bit of time.

I would like to help.  I work for HP Networking.

If you are having a problem with "losing access to network" using HP Networking MSM wireless products, I would like you to private message me on this forum. 

When you private message me, please provide full details of the problem that you're having, including the products, version of software, relevant log files, and, importantly, the support case number that you were provided when you reported the issue to HP Networking support.

I will get back to you.

Thank you.

--glen