Forum Discussion

Mauricio_Garant's avatar
Mauricio_Garant
Icon for Nimbostratus rankNimbostratus
Feb 06, 2015

http_process_state_prepend - Invalid action:0x109010

Hello Guys

 

I need your help. I have an LTM (Active/Passive). The Active unit remains with that error:

 

Feb 5 20:04:24 BOGEAWF5LTMP01 err tmm1[11181]: 011f0007:3: http_process_state_prepend - Invalid action:0x109010 (Server side: vip=/Common/VS_PX_IGUANA profile=http pool=/Common/Proxy_Pool server_ip=10.1.150.10)

 

I don´t have any irules in the Virtual Server.

 

That error appears randomly in /var/log/ltm.

 

Thanks

 

  • NikhilB_149913's avatar
    NikhilB_149913
    Historic F5 Account

    Seems like this may happen when mirroring takes place. I am not too familiar with this error but attached a url for your review.

     

    https://support.f5.com/kb/en-us/solutions/public/5000/900/sol5922.html

     

    If a virtual server is configured to mirror its connections to the standby unit, the standby BIG-IP system may log this error message. The error is most frequently seen when the virtual server is also configured to manipulate the server response. For example, when OneConnect, HTTP pipelining, compression, or caching are enabled in the HTTP profile.

     

    • Mauricio_Garant's avatar
      Mauricio_Garant
      Icon for Nimbostratus rankNimbostratus
      Thanks for your help NikhilB - The virtual server is not configured to mirror connections to stand by. I did apply the irule that i found in this link https://devcentral.f5.com/s/feed/0D51T00006i7YhuSAE but the "http_process_state_prepend - Invalid action:0x109010" still appears.
  • Seems like this may happen when mirroring takes place. I am not too familiar with this error but attached a url for your review.

     

    https://support.f5.com/kb/en-us/solutions/public/5000/900/sol5922.html

     

    If a virtual server is configured to mirror its connections to the standby unit, the standby BIG-IP system may log this error message. The error is most frequently seen when the virtual server is also configured to manipulate the server response. For example, when OneConnect, HTTP pipelining, compression, or caching are enabled in the HTTP profile.

     

  • shaggy's avatar
    shaggy
    Icon for Nimbostratus rankNimbostratus

    i have this log message show up often for one of my application configurations. i was able to track it down to a malformed HTTP response from the pool member. in my case, the pool member is sending an extra \r\n within the http response for a particular webpage. the f5 fixes and forwards the response to the user (i would guess this depends on how the response is malformed), but still logs the message.

     

    i used tcpdump, wireshark, and curl to track down the URL, request type, and issue with the response

     

    • Mauricio_Garant's avatar
      Mauricio_Garant
      Icon for Nimbostratus rankNimbostratus
      Thanks for your help Shaggy - I´m using the The virtual server for balancing an Explicit Proxys Bluecoat for Internet Content Filtering and I did apply the irule that i found in this link https://devcentral.f5.com/questions/issue-with-ltm-profile-http-after-upgrade-to-1151hf3 but the "http_process_state_prepend - Invalid action:0x109010" still appears. It´s very difficult use tcpdump for the thousand connections generated to/from proxies for found the Response that cause that event, due to the event occurs randomly.
  • i have this log message show up often for one of my application configurations. i was able to track it down to a malformed HTTP response from the pool member. in my case, the pool member is sending an extra \r\n within the http response for a particular webpage. the f5 fixes and forwards the response to the user (i would guess this depends on how the response is malformed), but still logs the message.

     

    i used tcpdump, wireshark, and curl to track down the URL, request type, and issue with the response

     

    • Mauricio_Garant's avatar
      Mauricio_Garant
      Icon for Nimbostratus rankNimbostratus
      Thanks for your help Shaggy - I´m using the The virtual server for balancing an Explicit Proxys Bluecoat for Internet Content Filtering and I did apply the irule that i found in this link https://devcentral.f5.com/s/feed/0D51T00006i7YhuSAE but the "http_process_state_prepend - Invalid action:0x109010" still appears. It´s very difficult use tcpdump for the thousand connections generated to/from proxies for found the Response that cause that event, due to the event occurs randomly.
  • Hi,

     

    http_process_state_prepend - Invalid action EV_INGRESS_DATA during ST_HTTP_PREPEND_HEADERS

     

    I am also getting the similar error in LTM logs and worried if this kind of error will impact to the business traffic or will make Pool Down or will make Virtual Server down. ?

     

    Pool Member have Microsoft-IIS/7.5 running with port 443.

     

    if anybody can help / share the experience on the same then would be great. this, as I don't configured any iRule, it is only simple iApp virtual server configured.

     

    regards / Neeraj J.

     

  • i had a similar issue. What fixed my issue was to change the response chuncking in the HTTP profile

     

  • Hi Joelc,

     

    I have seen the " Request Chunking" parameter has default is -:Preserve:

     

    what value you have selected which worked ? and any logic behind it.

     

    Rgds.//

     

    Neeraj Jagetia

     

  • The same problem. An VS in performance L4 for bypass an webservice https connection to the nodes. Using the http profile without any change and still looking for the problem (in my case the connections are reset in 80% aprox.). We are looking for the problem but any help is welcome too.

     

  • I ran into this error. This may not apply to everyone's scenario, but thought I would post our solution. It turned out to be something really simple. We were trying to do an SSL::disable serverside, which was working just fine. Problem is, "port translation" was not enabled on the VIP, so it was disabling SSL, but it was not sending the traffic to port 80 on the server side (even though the pool was defined as such). Enabling port translation on the VIP allowed it to correctly send the traffic to 80 after the SSL::disable serverside command and eliminate the error.