Losing access to network

[shayatin]

Hi, I have an MSM710, 10 APs, and about 80 users. 

I have a very odd situation.

A user can connect to an AP fine, but after a while (anywhere from 10 minutes to 4 hours) they will lose access to all network resources. No internet, no outlook, no TS, no DHCP, nothing.

Both the connection information on a laptop, and the connection information on the controller show the user as still connected. So they are not actually losing the wireless connection, just the network connection.

This happens to all users at the same time. The wired users are not affected.

Wireless users are authenticated by NPS / RADIUS. If a user attempts to reset their connection by disable / enable of their NIC, they will not receive an address from DHCP (ipconfig will show 169.254.x.x), but DHCP logs will show a succesful lease just granted. Also the RADIUS logs show recent authentication.

I'm very confused.

Thanks in advance.

[ManiSundaram]

Hi

  I also experience the same problem. Suddenly all wireless users traffic get slow / no traffic but the ssid is connected. I am using 5.5.1firmware with MSM710 and 4 MSM410 AP.

Thanks

Mani Sundaram

[Dr-Q]

Same issue here. Flash the AP to V5.5.2.1_MR_V5.5.2.1_MSM422.cim. Client lost internet connection but AP web management shown client is connected. HP Support, please solve this asap.  

[BCS-Tech]

We appear to be experiencing a similar situation.  We have a MSM760 Controller with 76 AP's (MSM410 & MSM460) across our compus.  Our users remain connected to the SSID but are unable to ping any resources on the network.  For us, disconnecting and reconnecting the client resolves the issue for a while, but it continues to happen.  I have not been able to pin-point a direct cause to the drop, it appears random.  Please advise.

[BGraham_1]

We are also experiencing similar issues since we upgraded to 5.5.1 in April. We have 2 locations, each with a MSM765zl managing 55-75 APs.We are usig 802.1x authenticating for managing WPA2 Enterprise key management. Clients will connect and then some will show up with a self-assigned ip (169.254.x.x) after a while. It seems to localized to a few APs, but not eveery client on the AP has the problem. If I restart the AP, it clears the problem up for awhile. It seems to be a problem with APs that have the most traffic.

A friend of mine hs a similar situation with 2 different sites without any special authentication, just using WPA2-PSK.

[MortenM]

Hi,

A client of mine had the same issues with clients disconnecting randomly, and getting 169.x.x.x addresses. They are using msm 760 and 12 access points (430). They have both dot1x and mac authenticated users.

I created a support case with HP, and got access to  5.5.1.4 software. That seemed to solve the issue regarding the 169.x.x.x addresses, but there still are some clients disconnecting randomly.

[JvR_NM]

Hello,

We have the same issues with a MSM760, with 802.1x users. Sometimes losing network connection while the ssid is still connected. This is very random, sometimes my collegea loses connection while I'm still connected (same AP). I think the problems started with firmware 5.5.1.0 (old version we uesed was 5.4.1.0). Going back to this firmware loses a lot of new settings, so is not an good option.

I found some errors in the MSM760 every time a connection is lost:

Sep  7 17:22:09 err eapolserver  SG932ZV000 Error while removing vlan id:50 (RADIUS) (mac-address='00-74-00-2F-C2-AA').
Sep  7 17:22:09 err eapolserver  SG932ZV000 [vlan] <__CONFIG_UnRegisterVLAN>: VLAN (ID='1-4094', interface='pvw0') not found in VLAN table.
Sep  7 16:44:11 err eapolserver  SG932ZV000 Error while removing vlan id:50 (RADIUS) (mac-address='00-00-F0-03-93-92').
Sep  7 16:44:11 err eapolserver  SG932ZV000 [vlan] <__CONFIG_UnRegisterVLAN>: VLAN (ID='1-4094', interface='pvw0') not found in VLAN table.

Does anyone has the same erros ?

[bkilgore]

I have been also been told about this problem multiple times by one of my customers.  They are running a MSM460 with about 50 MSM410 APs.  I don't see the problem when I'm on site because it is so random.  We did a wipe and reload of the controller in May when school was out but now being told it is still happening.  We upgraded the controller to 5.4.2.61-01-9824 per HP support.  I'm concerned on why you guys are seeing this issue in higher versions.  I was thinking it was a client problem but not sure if that is true now.  Has anyone had any luck with this since ?

[MortenM]

We are also seeing these errors:

Sep  9 11:02:35 err    eapolserver  CN16DWZ0GW Error while removing vlan id:14 (RADIUS) (mac-address='A0-88-B4-36-95-40').
Sep  9 11:02:35 err    eapolserver  CN16DWZ0GW [vlan] <__CONFIG_UnRegisterVLAN>: VLAN (ID='1-4094', interface='r2v1') not found in VLAN table.

Another noticable error we get is:

Sep  9 10:38:29 warning    eapolserver  CN16DWZ02T Unable to update Interim Traffic (mac-address='A0-88-B4-15-21-D8').

Software version on msm760: 5.5.2.14-01-10104

[tschaps]

Seeing similar issue here. I am just beginning a school rollout of about 80 MSM430's and 466's, as part of a full infrastructure upgrade, eventually with redundant controllers, but right now, I just have two configured in autonomous mode, four SSID's on three VLANs. Twice now we've had a room full of students working, and suddenly most of them have no network access -- the wifi is still connected, strong signal, IP address OK, but no IP traffic going through. It doesn't seem to 'recover,' either, until client is rebooted or put into sleep/standby mode and awakened.

HP, our admin will tell me to send back this $100k of HP equip and get Cisco back in here if this isn't resolved. Our vendor's engineers are coming tomorrow, and I'll let them deal with HP Support, but I am not sure the problem will happen when they are here! I don't think it's the network infrastructure, because the main MSM430 I've been tracking replaced an HP v-M200 configured the same way, so the switch port VLAN settings needed no changes, and we never had this issue with the V-M200.

I hope HP is paying attention to this one before it blows up.  

[ISoliman]

Please contact the support team, and also make sure the "wireless security filter" under teh VSC config is disabled "unchecked", you might need to resync the APs after that.

If same issue exists after that then support team will be able to help after providing the data/logs from the controller.

Kind Regards,

Islam

[tschaps]

thanks for the reply - if you were referring to me, the three I am testing are in autonomous mode, no controller yet, because it's backordered for another month. And I checked, the wireless security filter is unchecked on all of them. I am bummed, because I wanted to get a headstart installing these before the controller arrives and just setting them up individually. I have almost 70 left to install, but that is at a standstill until this gets resolved. thx

[dday0002]

I am having this same issue on software version 5.5.2.14 has anyone found a solution yet?

[BGraham_1]

I still have a case open with support from April. They ask for more information once in a while, but I am still not sure what they are finding. I have 150 more APs to deploy in November and I don't have a warm fuzzy feeling about it. Especially since they are MSM460s and I can't revert back to 5.4.x or 5.3.x.

[sjordet]

And I have the same problem. Have raised a case with support, but they asked for a lot of details, which I gave them. That's a month ago now... I have around 100 APs deployed, and plans for around 50 more...

[Piotras]

I had the same problem. You can try increase "retry interval" in the configuration of RADIUS.In my case it worked

[shayatin]

I am amazed to see all of these posts. I posted this months ago, and happened to be on here looking for a solution to a probably unrelated issue.

As way of an update, the reason I am so surprised is because I have gone round and round with HP support over these months, and HP has consistently told me that they have never heard another customer mention a similar issue. They had gotten me to the point of really thinking this was my issue, and in fact they have taken NO ownership of it. Meanwhile our company's wireless is notorious and has become the bane of my work week.

The problem still remains. It does not happen with any regularity. It has gone for 2 weeks without a hiccup and then bam it goes down hard. Then after a controller reboot it'll be stable again for a week or two.

Ultimately I feel like this HP solution is just not robust enough to support my users. I have already spent more in troubleshooting than the entire purchase price. We will have to replace this with Cisco or live with the unreliability.

Thank you all for the input, I'm sorry we are living with an inferior solution.

[sjordet]

Hi!

I eventually, after several email to ask if they were dead, got an answer from support again. They said that they think my problem is solved with 5.5.3. (Something about auto-channel bug).

I installed this last weekend, and haven't really had time to test properly yet, but the initial observations seems positive... Anyone else have any experience about 5.5.3?

Regards,

Stian

[cenk sasmaztin]

this issue fixed 5.5.3 software

https://h10145.www1.hp.com/downloads/SoftwareReleases.aspx?ProductNumber=J9420A

Beamforming

(Only supported on the E-MSM430, E-MSM460, E-MSM466.)

The beamforming feature is available in v5.5.1 and higher for all product versions as follows:

Fixes

The following issues have been fixed since release 5.5.2.0:

Model WW Americas Japan Israel

E-MSM430 J9651A J9650A J9652A J9653A

E-MSM460 J9591A J9590A J9589A J9618A

E-MSM466 J9622A J9621A J9620A

ID Description

55920 When the auto channel feature is enabled on the radio page, and the channel exclusion

list is set to 1, 6, and 11, most APs incorrectly end up on the same channel.

55778 The auto channel feature fails to select non-overlapping channels in the 2.4GHz band.

53851 HTML-based authentication only works on the team manager. Team members fail to

provide HTML authentication. Wireless clients receive the login web page but receive a

blank page after providing login credentials.

53254 Mobility Traffic Manager (MTM) clients lose IP connectivity when the egress network is

configured to operate over a range of VLANs.

53040 The error message: "Generic Error. (255)" is displayed if an empty username or

password is specified when clicking the

Join Realm Now button on the Controller >>

Authentication > Active Directory

page. A more descriptive error message should be

displayed indicating that the username provided does not exist on Active Directory or is

ambiguous.

52997 When configuring a VSC, the error message: "When identification of stations is based

on IP address only, MAC-based authentication methods cannot be selected." appears

only when

Identify stations based on IP address and HTML Authentication options

are enabled. It should also appear when any other authentication method is enabled.

52915 In a teaming environment with the location-aware feature enabled the CPU utilization

goes very high and this syslog message appears:

SendMessageToListeners: Error sending packet: Resource temporarily

unavailable(11) to listener #5(/tmp/busclient.iappd_sc.458.1)

44471 The

Wireless > Neighborhood page displays duplicate and erroneous information.

44130 The

Wireless > Neighborhood page always displays WPA2 regardless of the

configuration of neighboring devices.

44083 Updating a Certificate Revocation List (CRL) when an invalid certificate is selected

causes the management tool to hang.

43962 (Only E-MSM430, E-MSM460, and E-MSM466.) The beacon/probe shows an incorrect

vendor ID.

[BGraham_1]

I see many things fixed in the release notes, but I do not see that this resolves the issue originall posted by shayatin on 4-11-2011. Am I  missng something?

[cenk sasmaztin]

please send me show config print msm controller

[cenk sasmaztin]

I can test new software version this weekend

[BGraham_1]

I would send you our config, but we have a fairly complex setup with 12 VSCs (some with 802.1X, some with WPA2) and as many VLANs to match.

I have a current case open with TAC, but I have been following this thread closely due to the similarity to the issues we are having.

My question a few minutes ago is because you said "this issue fixed 5.5.3 software". but when I review the release notes, I saw nothing conclusive regarding the issue that started this thread. Two "Fixes" that look like they might be related are:

43253 When a large number of wireless clients are roaming between APs or disconnecting/
reconnecting (for example, in a busy public environment) the maximum number of
supported wireless clients slowly decreases because internal management cannot keep
up with the changes.

43478 Egress VLANs in a VSC binding are not synchronized properly on all managed APs.

Although I remain hopeful this will be resolved, I think that saying 5.5.3 is the resolution is still a bit premature.

I will post what HP support finds with our case when resolved, so that others may try it in their network to see if it works in their situation.

Kind Regards,

Bob

[cenk sasmaztin]

I can setup last week 70 310 ap and MSM 760 controller
seven ssid and eggress vlan config one ssid 802.1x ,four ssid wpa and one ssid html authentication.

My msm 760 software version 5.5.1
can't see any eror wireless user
still working without a complaint

I think your issue 4XX access points and fix this issue 5.5.3 software this is no controller issue

Re: Losing access to networkOptions
Edit Reply

Mark as New

Bookmark

Subscribe

Subscribe to RSS Feed

Highlight

Print

Email to a Friend

Report Inappropriate Content

an hour ago

this issue fixed with 5.5.3 software



https://h10145.www1.hp.com/downloads/SoftwareReleases.aspx?ProductNumber=J9420A



Beamforming

(Only supported on the E-MSM430, E-MSM460, E-MSM466.)

The beamforming feature is available in v5.5.1 and higher for all product versions as follows:

Fixes

The following issues have been fixed since release 5.5.2.0:

Model WW Americas Japan Israel

E-MSM430 J9651A J9650A J9652A J9653A

E-MSM460 J9591A J9590A J9589A J9618A

E-MSM466 J9622A J9621A J9620A

[sjordet]

What my contact claimed would fix it was this:

Defect 102596 (Auto channel fix for MSM46x & MSM430)

Defect 102601 (Auto channel fix for MSM422 & MSM410... (legacy etc...) )"