Forum Discussion
The error is due to the server-side HTTP filter seeing ingress when it doesn't expect it.
This can happen if the content length of the response is wrong. It can also happen if multiple responses are triggered by a single request. It can happen if the protocol isn't actually HTTP, and the server isn't following the standard request-response paradigm. Finally, this can happen in a HA setup if the master and standby get out of sync. (The standby can get data it doesn't expect yet.)
The only way to be sure is to use a TCP dump and look at the traffic.
This condition will not mark a pool down, or take down a virtual server. It just says that the HTTP filter is rejecting that particular connection because it doesn't understand it any more.
In an ADC scenario, this error typically indicates that there is some configuration error in the back-end web servers. In a service provider scenario, this indicates that you probably should be using the "Transparent HTTP" profile instead.
- Neeraj_Jags_152Sep 23, 2015CirrusHi sfuerst, thanks, So can I understand this as per below. this is generally a intermittent (rarely) scenario. so whenever this will be observed by LTM then only it will impact to the specific Client session (not the all client same time) . so most of people will feel services is working ... !!! -and also after tcpdump what to search in the capture, exactly to see if this issue is in Nth frame and to find what exact , which caused the problem , how , if you were able to dig so ? Regards / Neeraj Jagetia