[GitHub] trafficserver issue #1412: Assertion in HttpSM::tunnel_handler_post
Github user oknet commented on the issue: https://github.com/apache/trafficserver/issues/1412 PR#1583 merged to fix the issue. --- If your project is set up for it, you can reply to this email and have your reply appear on GitHub as well. If your project does not have this feature enabled and wishes so, or if the feature is enabled but not working, please contact infrastructure at infrastruct...@apache.org or file a JIRA ticket with INFRA. ---
[GitHub] trafficserver issue #1412: Assertion in HttpSM::tunnel_handler_post
Github user oknet commented on the issue: https://github.com/apache/trafficserver/issues/1412 @zwoop Could you have a look on PR #1583 ? I think it is useful for this issue. But I still need your core file to confirm it. --- If your project is set up for it, you can reply to this email and have your reply appear on GitHub as well. If your project does not have this feature enabled and wishes so, or if the feature is enabled but not working, please contact infrastructure at infrastruct...@apache.org or file a JIRA ticket with INFRA. ---
[GitHub] trafficserver issue #1412: Assertion in HttpSM::tunnel_handler_post
Github user oknet commented on the issue:
https://github.com/apache/trafficserver/issues/1412
Maybe I found the reason:
```
3480 int
3481 HttpSM::tunnel_handler_post_ua(int event, HttpTunnelProducer *p)
3482 {
3483 STATE_ENTER(::tunnel_handler_post_ua, event);
3484 client_request_body_bytes = p->init_bytes_done + p->bytes_read;
3485 int64_t alloc_index, nbytes;
3486 IOBufferReader *buf_start;
3487
3488 switch (event) {
3489 case VC_EVENT_INACTIVITY_TIMEOUT:
3490 case VC_EVENT_ACTIVE_TIMEOUT:
...
3520 p->vc->do_io_write(this, nbytes, buf_start);
3521 p->vc->do_io_shutdown(IO_SHUTDOWN_READ);
3522 return 0; // > we do not reset vc's timeout
3523 }
```
InactivityCop will callback again in the next second and the default httpsm
handler is tunnel_handler_post().
I think we should handle TIMEOUT event in tunnel_handler_post() since the
timeout event get from ua_session.
I will have a details analysis on it and create a PR to fix this issue.
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at infrastruct...@apache.org or file a JIRA ticket
with INFRA.
---
[GitHub] trafficserver issue #1412: Assertion in HttpSM::tunnel_handler_post
Github user zwoop commented on the issue: https://github.com/apache/trafficserver/issues/1412 I'll see if I can get a core out of this again, for some reason, the box isn't producing core files :/ --- If your project is set up for it, you can reply to this email and have your reply appear on GitHub as well. If your project does not have this feature enabled and wishes so, or if the feature is enabled but not working, please contact infrastructure at infrastruct...@apache.org or file a JIRA ticket with INFRA. ---
[GitHub] trafficserver issue #1412: Assertion in HttpSM::tunnel_handler_post
Github user zwoop commented on the issue: https://github.com/apache/trafficserver/issues/1412 We're still seeing this on Docs, with the current 7.1.x and all the backports. --- If your project is set up for it, you can reply to this email and have your reply appear on GitHub as well. If your project does not have this feature enabled and wishes so, or if the feature is enabled but not working, please contact infrastructure at infrastruct...@apache.org or file a JIRA ticket with INFRA. ---
[GitHub] trafficserver issue #1412: Assertion in HttpSM::tunnel_handler_post
Github user oknet commented on the issue: https://github.com/apache/trafficserver/issues/1412 @zwoop can u show me the HttpSM.history ? --- If your project is set up for it, you can reply to this email and have your reply appear on GitHub as well. If your project does not have this feature enabled and wishes so, or if the feature is enabled but not working, please contact infrastructure at infrastruct...@apache.org or file a JIRA ticket with INFRA. ---
[GitHub] trafficserver issue #1412: Assertion in HttpSM::tunnel_handler_post
Github user zwoop commented on the issue: https://github.com/apache/trafficserver/issues/1412 @oknet Did any of your commits address this issue? --- If your project is set up for it, you can reply to this email and have your reply appear on GitHub as well. If your project does not have this feature enabled and wishes so, or if the feature is enabled but not working, please contact infrastructure at infrastruct...@apache.org or file a JIRA ticket with INFRA. ---
[GitHub] trafficserver issue #1412: Assertion in HttpSM::tunnel_handler_post
Github user oknet commented on the issue: https://github.com/apache/trafficserver/issues/1412 and HttpSM::history is needed. --- If your project is set up for it, you can reply to this email and have your reply appear on GitHub as well. If your project does not have this feature enabled and wishes so, or if the feature is enabled but not working, please contact infrastructure at infrastruct...@apache.org or file a JIRA ticket with INFRA. ---
[GitHub] trafficserver issue #1412: Assertion in HttpSM::tunnel_handler_post
Github user oknet commented on the issue: https://github.com/apache/trafficserver/issues/1412 Can you show me the netvc_context of the UnixNetVConnection ? in order to identify the side of it (client vc or server vc ?) --- If your project is set up for it, you can reply to this email and have your reply appear on GitHub as well. If your project does not have this feature enabled and wishes so, or if the feature is enabled but not working, please contact infrastructure at infrastruct...@apache.org or file a JIRA ticket with INFRA. ---
[GitHub] trafficserver issue #1412: Assertion in HttpSM::tunnel_handler_post
Github user oknet commented on the issue: https://github.com/apache/trafficserver/issues/1412 I found a bug related to this issue: HttpSM::state_request_wait_for_transform_read could receive the VC_EVENT_EOS, VC_EVENT_ERROR from TransformTerminus::handle_event and transfer these 2 events to HttpSM::tunnel_handler_post. but the p->handler_state is not (may be) set to HTTP_SM_POST_UA_FAIL. I will create a pr later. --- If your project is set up for it, you can reply to this email and have your reply appear on GitHub as well. If your project does not have this feature enabled and wishes so, or if the feature is enabled but not working, please contact infrastructure at infrastruct...@apache.org or file a JIRA ticket with INFRA. ---
[GitHub] trafficserver issue #1412: Assertion in HttpSM::tunnel_handler_post
Github user bryancall commented on the issue: https://github.com/apache/trafficserver/issues/1412 The event is VC_EVENT_INACTIVITY_TIMEOUT (value 105) when it asserts. --- If your project is set up for it, you can reply to this email and have your reply appear on GitHub as well. If your project does not have this feature enabled and wishes so, or if the feature is enabled but not working, please contact infrastructure at infrastruct...@apache.org or file a JIRA ticket with INFRA. ---
[GitHub] trafficserver issue #1412: Assertion in HttpSM::tunnel_handler_post
GitHub user zwoop opened an issue: https://github.com/apache/trafficserver/issues/1412 Assertion in HttpSM::tunnel_handler_post I think we've seen this before, but filing this again to make sure I don't lose the info: ``` (gdb) bt #0 0x74fed1d7 in __GI_raise (sig=sig@entry=6) at ../nptl/sysdeps/unix/sysv/linux/raise.c:56 #1 0x74fee8c8 in __GI_abort () at abort.c:90 #2 0x77bb89c8 in ink_abort (message_format=0x77bca590 "%s:%d: failed assertion `%s`") at ink_error.cc:99 #3 0x77bb609e in _ink_assert (expression=0x800058 "(event == VC_EVENT_WRITE_COMPLETE) || (event == VC_EVENT_EOS)", file=0x7fe78c "HttpSM.cc", line=2763) at ink_assert.cc:37 #4 0x005f58cd in HttpSM::tunnel_handler_post (this=0x2aaabfcbe8f0, event=105, data=0x2aaabf009620) at HttpSM.cc:2763 #5 0x005f534b in HttpSM::main_handler (this=0x2aaabfcbe8f0, event=105, data=0x2aaabf009620) at HttpSM.cc:2665 #6 0x0051a5c8 in Continuation::handleEvent (this=0x2aaabfcbe8f0, event=105, data=0x2aaabf009620) at /usr/local/src/trafficserver/iocore/eventsystem/I_Continuation.h:153 #7 0x007a8235 in write_signal_and_update (event=105, vc=0x2aaabf009480) at UnixNetVConnection.cc:176 #8 0x007ac11a in UnixNetVConnection::mainEvent (this=0x2aaabf009480, event=1, e=0x2b5feb10) at UnixNetVConnection.cc:1289 #9 0x0051a5c8 in Continuation::handleEvent (this=0x2aaabf009480, event=1, data=0x2b5feb10) at /usr/local/src/trafficserver/iocore/eventsystem/I_Continuation.h:153 #10 0x007a10b6 in NetHandler::_close_vc (this=0x72409e50, vc=0x2aaabf009480, now=1486031338941225970, handle_event=@0x2b5fec28: 0, closed=@0x2b5fec2c: 0, total_idle_time=@0x2b5fec24: 0, total_idle_count=@0x2b5fec20: 0) at UnixNet.cc:679 #11 0x007a09a6 in NetHandler::manage_active_queue (this=0x72409e50, ignore_queue_size=true) at UnixNet.cc:588 #12 0x007a1cbb in InactivityCop::check_inactivity (this=0x73ecd100, event=2, e=0x2bc05a00) at UnixNet.cc:94 #13 0x0051a5c8 in Continuation::handleEvent (this=0x73ecd100, event=2, data=0x2bc05a00) at /usr/local/src/trafficserver/iocore/eventsystem/I_Continuation.h:153 #14 0x007cc28c in EThread::process_event (this=0x72406000, e=0x2bc05a00, calling_code=2) at UnixEThread.cc:143 #15 0x007cc6c7 in EThread::execute (this=0x72406000) at UnixEThread.cc:225 #16 0x007cb890 in spawn_thread_internal (a=0x73ec9040) at Thread.cc:84 #17 0x75a9ddc5 in start_thread (arg=0x2b5ff700) at pthread_create.c:308 #18 0x750af73d in clone () at ../sysdeps/unix/sysv/linux/x86_64/clone.S:113 (gdb) frame 4 #4 0x005f58cd in HttpSM::tunnel_handler_post (this=0x2aaabfcbe8f0, event=105, data=0x2aaabf009620) at HttpSM.cc:2763 2763 ink_assert((event == VC_EVENT_WRITE_COMPLETE) || (event == VC_EVENT_EOS)); (gdb) print event $1 = 105 ``` --- If your project is set up for it, you can reply to this email and have your reply appear on GitHub as well. If your project does not have this feature enabled and wishes so, or if the feature is enabled but not working, please contact infrastructure at infrastruct...@apache.org or file a JIRA ticket with INFRA. ---
