Forum Discussion

Terrence's avatar
Terrence
Icon for Nimbostratus rankNimbostratus
Mar 21, 2013

Odd Connection resets returned by

Hello,

 

 

We have followed the design guide to deploy exchange behind our edge gateways and ltms.

 

Couple of facts about our environment

 

LTMS 3600s running BIG-IP 11.1.0 Build 2027.0 Hotfix HF2

 

Edge Gateways 1600s running BIG-IP 11.2.0 Build 2713.0 Hotfix HF3

 

Nothing is Natted(CAS nodes see real ips)

 

It is deployed exactly as specified by the design guide.

 

 

Basically what is happening at random. When a user logs in via the Edge gateway to owa, randomly the ltm will return a rst. Upon refreshing the browser traffic flows as normal.

 

 

I have confirmed that the reset is being returned from the ltm, as I followed the traffic back doing packet captures. Comparing the packet captures from a regular request and a request that recieves the reset, they appear to exactly the same request.

 

 

Is there anything else that I should look at? I asked here, because I dont necessarily believe it is directly related to the Exchange iApp or Design guide, just wondering what else could cause resets to be randomly returned.

 

 

I know this post lacks a lot of information, I just wasnt sure what information to provide, or the type of response I would get.

 

 

 

 

7 Replies

  • have you tried to log reset cause?

     

     

    sol13223: Configuring the BIG-IP system to log TCP RST packets

     

    http://support.f5.com/kb/en-us/solutions/public/13000/200/sol13223.html
  • Thank you nitass, that solution was very helpful, however the messages that I recieved were less than insightful.

     

     

    The solution to log the packets unfortunately didnt work as it was quickly throttled by syslog.

     

     

    I added the reason to the pkt, and the reset cause for the traffic in question was

     

    "Reset cause: BIG-IP: Unknown reason"

     

     

    Is there known scenerios that I might reference as to why the ltm would return a reset with an unknown reason?

     

     

  • are you using match across setting in persistence profile? if so, can you open a support case to verify?

     

     

    ID385333 - (377421-1) Match-across-* persistence across TMMs requests record, then fails to match
  • this is askf5 sol.

     

     

    sol14061: Persistence may fail when a persistence profile is configured with Match Across Virtual Servers/Pools/Services

     

    http://support.f5.com/kb/en-us/solutions/public/14000/000/sol14061.html
  • Thank you. I think you might have hit the nail on the head with this one.

     

     

    We are using the aformentioned persistence, as it is extensively used in the Exchange iApp as mentioned.

     

     

    I am kind of thinking that I wont open a ticket, as they will just tell me to upgrade, which I am okay with doing anyways. If the problem persists after an upgrade, I will then open a ticket.

     

     

  • Terry,

     

    Did upgrading resolve the issue for you? If so, what version are you currently running?

     

    I am having a problem that sounds quite similar, but we are running on V11.3-HF7