Broadcom BASP Teaming

Hi,

We have two Dell servers (running Windows Server 2003 SP2) with two broadcom cards that are teamed into one "BASP Virtual Adapters"
However, we have problems with this teaming.

Sometimes, on these servers (let's say it has an IP of 192.168.1.55) can not been seen from an external subnet (let's say 192.168.2.88) : which is a problem since our VPN is built that way and none of our overseas users can not access the applications on this server anymore... It can still be seen from a machine in the same subnet (same office in which the servers are), though.

After some investigation, I discovered that visiting an external website (like google.com) makes the server visible again. It is weird. For now, I wrote a script doing something similar (wget www.google.com) as a scheduled task and launch it every few minutes as the problem keeps happening without any reason or recognizible pattern.

Really, something seems wrong : we received the machine this way from Dell and it happens on two servers, so it can not be an hardware issue with the network cards) : I already tried to upgrade the drivers, but the issue is still there. If I unteam one of the network cards, the problem is gone on the unteamed one but keeps happening on the teamed one.

Does anyone have experience with this Broadcom Teaming / BASP Virtual Adapter ?
It is the first time we actually use this system so maybe we missed something ?

If it can not be fixed, what is really the loss of this teaming anyway ? It sure is convenient to have a machine with several NIC referrable by a single IP address, but if we are to get rid of this teaming system, what would be the performance penalty ?

Thanks.
LVL 4
VorenusAsked:
Who is Participating?
 
grsteedCommented:
It sounds like it's losing the Default Gateway.  Some commands to check.

ipconfig /all   (see what it thinks the Gateway address is for that IP)
route print   (check to see where it sends other than local traffic)
arp -a          See what the entry for the gateway looks like.

Do these commands while it having problems and after you "fix" it and compare.

As far as the performance hit, it depends on the speed of the NIC's and the amount of traffic.  It would certainly be worth trying. See if it behaves the same way with a single NIC. It may or may not be a problem with the teaming.

Cheers,
Gary
0
 
VorenusAuthor Commented:
Hi grsteed,

First, sorry for the delay, but I had to wait for the problem to happen again, over which I have no control on :
Thanks, you are right : it was indeed losing its gateway.

When I do an ipconfig /all while the problem happens, the default gateway field is empty, and when I fix it ( going to a website), and redo an ipconfig /all after, it gets back its gateway as 192.168.1.1 ?

Routes are the same while and after the problem.

arp -a during the problem :


Interface: 10.1.1.50 --- 0x10004
  Internet Address      Physical Address      Type
  10.1.1.51             00-04-23-ce-7b-ef     dynamic  
  192.168.1.1           00-00-00-00-00-00     invalid  

Interface: 192.168.1.50 --- 0x10005
  Internet Address      Physical Address      Type
  192.168.1.1           00-a0-8e-7f-9a-2d     dynamic  
  192.168.1.8           00-18-8b-85-30-c6     dynamic  
  192.168.1.10          00-18-8b-84-a8-87     dynamic  
  192.168.1.51          00-04-23-ce-7b-ee     dynamic  
  192.168.1.55          00-19-b9-df-bc-03     dynamic  
  192.168.1.60          00-18-8b-85-30-ad     dynamic  
  192.168.1.80          00-19-b9-e0-23-31     dynamic  
  192.168.1.114         00-11-43-7b-19-07     dynamic  
  192.168.1.124         00-0b-db-ac-48-4d     dynamic  
  192.168.1.148         00-08-74-a6-91-8d     dynamic  

arp -a once the problem is fixed :

Interface: 10.1.1.50 --- 0x10004
  Internet Address      Physical Address      Type
  10.1.1.51             00-04-23-ce-7b-ef     dynamic  

Interface: 192.168.1.50 --- 0x10005
  Internet Address      Physical Address      Type
  192.168.1.1           00-a0-8e-7f-9a-2d     dynamic  
  192.168.1.8           00-18-8b-85-30-c6     dynamic  
  192.168.1.10          00-18-8b-84-a8-87     dynamic  
  192.168.1.51          00-04-23-ce-7b-ee     dynamic  
  192.168.1.55          00-19-b9-df-bc-03     dynamic  
  192.168.1.60          00-18-8b-85-30-ad     dynamic  
  192.168.1.80          00-19-b9-e0-23-31     dynamic  
  192.168.1.114         00-11-43-7b-19-07     dynamic  
  192.168.1.124         00-0b-db-ac-48-4d     dynamic  
  192.168.1.148         00-08-74-a6-91-8d     dynamic  

It looks like that when the problem happens : the gateway gets into to 10.1.1.50 (a NIC used for Windows cluster communication) as well, while when everything works fine, it shouldn't be.

Any idea why the 192.168.1.50 adapter would stop seeing the gateway and how to fix it once for all ?

Thanks.
0
 
VorenusAuthor Commented:
By the way, forgot to say that the NICs on this machine are statically configured and not using DHCP for any cards on this server and that 192.168.1.1 is properly set up in the gateway edit box on their properties.
0
On-Demand: Securing Your Wi-Fi for Summer Travel

Traveling this summer?Check out our on-demand webinar to learn about the importance of Wi-Fi security and 3 easy measures you can start taking immediately to protect your private data while using public Wi-Fi. Follow us today to learn more!

 
grsteedCommented:
That is unusual for that GW address to show up on the other NIC.  

Is bridging between those interfaces turned on in Network Connections > Advanced?  Does this server have routing turned on in RRAS?

Could those networks be connected together external to the server?  

Do you have multiple GW address configured on the server?  There should only be one per this article.

http://www.microsoft.com/technet/community/columns/cableguy/cg0903.mspx

"If you have multiple interfaces connected to different disjoint networks, you should not configure a default gateway for each interface. Disjoint networks are networks that do not provide symmetric reachability to each other at the Network layer. Symmetric reachability exists when packets can be sent directly to and received directly from arbitrary endpoints without address translation."

Cheers,
Gary
0
 
VorenusAuthor Commented:
Thanks, I think that it was because there was also a gateway address setup on another NIC of this server : the arp -a showed that when the problem occured, the gateway was also on that other NIC while it wasn't when everything was working on (only on the main NIC).

It may be a bit early to confirm the problem is gone for good but it has been two days now without it coming back again, so thanks !
0
 
grsteedCommented:
Glad to hear it. Hopefully it stays that way.

Cheers,
Gary
0
 
punkrawkdude99Commented:
it wasnt mentioned in any of the comments but this can also happen

for instance you could have setup lan adapter 1 with an ip gw and subnet and then later decide to add teaming ...the registry ends up getting all screwed up this way and you need to clean up the adapter settings


    From the Windows Registry Editor, navigate to:

    HKLM/System/CurrentControlSet/Services/Tcpip/Parameters/Interfaces/%CLSID%

Look at each adapter... the answer is probably there
0
Question has a verified solution.

Are you are experiencing a similar issue? Get a personalized answer when you ask a related question.

Have a better answer? Share it in a comment.

All Courses

From novice to tech pro — start learning today.