Details

    • Type: Bug Bug
    • Status: Closed
    • Priority: Blocker Blocker
    • Resolution: Fixed
    • Affects Version/s: None
    • Fix Version/s: 2.1.3
    • Component/s: Core
    • Labels:
      None

      Description

      Reverse proxy no longer works in ATS, it seems that it tries to proxy back to itself instead of the destination host. I see

      [Sep 3 12:36:17.745] Server

      {140386788771600} DEBUG: (http) [0] State Transition: CACHE_ISSUE_WRITE -> ORIGIN_SERVER_OPEN
      [Sep 3 12:36:17.745] Server {140386788771600}

      DEBUG: (http_track) entered inside do_http_server_open
      [Sep 3 12:36:17.745] Server

      {140386788771600} DEBUG: (http) [0] open connection to loki.ogre.com: 192.168.201.14
      [Sep 3 12:36:17.745] Server {140386788771600}

      DEBUG: (http_seq) [HttpSM::do_http_server_open] Sending request to server
      [Sep 3 12:36:17.745] Server

      {140386788771600} DEBUG: (http) calling netProcessor.connect_re
      [Sep 3 12:36:17.745] Server {140386788771600}

      DEBUG: (http) [0] [HttpSM::main_handler, NET_EVENT_OPEN]
      [Sep 3 12:36:17.745] Server

      {140386726496016}

      DEBUG: (http_tproxy) Marking accepted connect on 32f81e0 as not outbound transparent.
      [Sep 3 12:36:17.745] Server

      {140386788771600} DEBUG: (http_track) entered inside state_http_server_open
      [Sep 3 12:36:17.745] Server {140386788771600}

      DEBUG: (http) [0] [&HttpSM::state_http_server_open, NET_EVENT_OPEN]

      when I expect something like the below (the origin server, RHS of remap.config rule, is www.theorigin.com):

      [Sep 3 12:37:55.805] Server

      {140243282487056} DEBUG: (http) [0] State Transition: CACHE_ISSUE_WRITE -> ORIGIN_SERVER_OPEN
      [Sep 3 12:37:55.805] Server {140243282487056}

      DEBUG: (http_track) entered inside do_http_server_open
      [Sep 3 12:37:55.805] Server

      {140243282487056} DEBUG: (http) [0] open connection to l.yimg.com: 68.142.250.160
      [Sep 3 12:37:55.805] Server {140243282487056}

      DEBUG: (http_seq) [HttpSM::do_http_server_open] Sending request to server
      [Sep 3 12:37:55.805] Server

      {140243282487056} DEBUG: (http) calling netProcessor.connect_re
      [Sep 3 12:37:55.842] Server {140243282487056}

      DEBUG: (http) [0] [HttpSM::main_handler, NET_EVENT_OPEN]
      [Sep 3 12:37:55.842] Server

      {140243282487056} DEBUG: (http_track) entered inside state_http_server_open
      [Sep 3 12:37:55.842] Server {140243282487056}

      DEBUG: (http) [0] [&HttpSM::state_http_server_open, NET_EVENT_OPEN]
      [Sep 3 12:37:55.842] Server

      {140243282487056} DEBUG: (http_ss) [0] session born, netvc 0x7f8cd8012f40
      [Sep 3 12:37:55.842] Server {140243282487056}

      DEBUG: (http) [0] [HttpSM::main_handler, VC_EVENT_WRITE_COMPLETE]
      [Sep 3 12:37:55.842] Server

      {140243282487056}

      DEBUG: (http) [0] [&HttpSM::state_send_server_request_header, VC_EVENT_WRITE_COMPLETE]

        Activity

        Gavin made changes -
        Workflow jira [ 12519694 ] TS Workflow [ 12522702 ]
        Alan M. Carroll made changes -
        Status Resolved [ 5 ] Closed [ 6 ]
        Hide
        Alan M. Carroll added a comment -

        Patch active for several days, no further issues.

        Show
        Alan M. Carroll added a comment - Patch active for several days, no further issues.
        Alan M. Carroll made changes -
        Status Open [ 1 ] Resolved [ 5 ]
        Resolution Fixed [ 1 ]
        Hide
        Alan M. Carroll added a comment -

        Patched checked in.

        Show
        Alan M. Carroll added a comment - Patched checked in.
        Hide
        Leif Hedstrom added a comment -

        Alan: I think we can close this one, right ?

        Show
        Leif Hedstrom added a comment - Alan: I think we can close this one, right ?
        Hide
        Alan M. Carroll added a comment -

        Updated fix submitted r992525. Moved marking the HTTP header cache dirty from its use to RemapProcess where the header is actually changed.

        Show
        Alan M. Carroll added a comment - Updated fix submitted r992525. Moved marking the HTTP header cache dirty from its use to RemapProcess where the header is actually changed.
        Hide
        Alan M. Carroll added a comment -

        I checked in a quick and dirty patch. I am leaving the bug open because I don't like the fix, it's just a work around. I need to find the real problem and fix that. It's clear, however, that the problem lies in the host name caching logic I added to the HTTPHdr object. It is not getting updated / marked dirty after remapping.

        Show
        Alan M. Carroll added a comment - I checked in a quick and dirty patch. I am leaving the bug open because I don't like the fix, it's just a work around. I need to find the real problem and fix that. It's clear, however, that the problem lies in the host name caching logic I added to the HTTPHdr object. It is not getting updated / marked dirty after remapping.
        Alan M. Carroll made changes -
        Field Original Value New Value
        Assignee Alan M. Carroll [ amc ]
        Leif Hedstrom created issue -

          People

          • Assignee:
            Alan M. Carroll
            Reporter:
            Leif Hedstrom
          • Votes:
            0 Vote for this issue
            Watchers:
            0 Start watching this issue

            Dates

            • Created:
              Updated:
              Resolved:

              Development