You are viewing a plain text version of this content. The canonical link for it is here.

Posted to commits@pulsar.apache.org by GitBox <gi...@apache.org> on 2020/07/16 22:14:38 UTC

[GitHub] [pulsar] klwilson227 opened a new issue #7565: PulsarClient-cpp segv after close() method is called.

klwilson227 opened a new issue #7565:
URL: https://github.com/apache/pulsar/issues/7565


   **Describe the bug**
   After the execution of the close method, a segmentation violation can be seen from time to time when either the keepAliveTimer or the consumerStatsTimer pops during or soon after the close operation. The segv shows the following stack trace:
   
   Signal: [0x000000000000000b] PID: [0x0000000000048988] PC: [0x00007f02e615c1b6] FP: [0x00007f02c40d0180] SIGSEGV: SEGV_MAPERR SI_ADDR : [0x0000000000000000]
   (_ZN6Basics9Backtrace11DoBacktraceEiiPvS1_+0x8f1) [0x4074ca1]
   (_ZN6Basics20GlobalSignalHandlers14logFatalSignalEiPvS1_+0xd2) [0x40c91b2]
   (_ZN6BasicsL26sigHandler_withinATryCatchEiP7siginfoPv+0x2c2) [0x40ca6b2]
   (_ZN6BasicsL10sigHandlerEiP7siginfoPv+0xe) [0x40ca77e]
   (_L_unlock_13+0x34) [0x7f03dc81e5d0]
   (_ZN5boost4asio20basic_deadline_timerINS_10posix_time5ptimeENS0_11time_traitsIS3_EENS0_22deadline_timer_serviceIS3_S5_EEE16expires_from_nowERKNS2_13time_durationE+0x16) [0x7f02e615c1b6]
   (UNKNOWN) [0x7f02e61d3768]
   (UNKNOWN) [0x7f02e61d3de3]
   (_ZN5boost4asio3ssl7contextD1Ev+0xdc1) [0x7f02e61e57b1]
   (_ZN5boost4asio6detail15task_io_service3runERNS_6system10error_codeE+0x311) [0x7f02e6121331]
   (_ZNSt14_Function_base13_Base_managerISt5_BindIFSt7_Mem_fnIMN6pulsar12ConsumerImplEFvNS3_6ResultENS3_9MessageIdESt8functionIFvS5_S6_EEEESt10shared_ptrIS4_ESt12_PlaceholderILi1EESF_ILi2EES9_EEE10_M_managerERSt9_Any_dataRKSL_St18_Manager_operation+0xaf6) [0x7f02e611e2c6]
   (_ZNSt6thread5_ImplISt12_Bind_simpleIFSt5_BindIFSt7_Mem_fnIMN6pulsar15ExecutorServiceEFvSt10shared_ptrIN5boost4asio10io_serviceEEEEPS5_SA_EEvEEE6_M_runEv+0x52) [0x7f02e6122772]
   (_ZNSt11this_thread11__sleep_forENSt6chrono8durationIlSt5ratioILl1ELl1EEEENS1_IlS2_ILl1ELl1000000000EEEE+0x1c0) [0x7f03dcde2070]
   (start_thread+0xc5) [0x7f03dc816dd5]
   (clone+0x6d) [0x7f03dc13402d]
   
   Steps to reproduce the behavior:
   This problem is difficult to reproduce without tweaking the source base. As this exception is based on timing. 
   To stress the environment add a short sleep to the ClientConnection::close() just after the mutex has been granted. This will give a long window for all of the timers to pop and align on the mutex if necessary. 
   
   **Expected behavior**
   Even with the sleep in place the timers should pop and be handled. So that they do not cause a segmentation violation and bring down the application for which the client is embedded with. 
   
   **Screenshots**
   N/A
   
   **Desktop (please complete the following information):**
    - OS: Centos
   
   **Additional context**
   Add any other context about the problem here.
   


----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

For queries about this service, please contact Infrastructure at:
users@infra.apache.org

[GitHub] [pulsar] merlimat commented on issue #7565: PulsarClient-cpp segv after close() method is called.

Posted by GitBox <gi...@apache.org>.

merlimat commented on issue #7565:
URL: https://github.com/apache/pulsar/issues/7565#issuecomment-659741695


   @klwilson227 Do you have a simple way to reproduce this (even if it involves putting `sleep()` calls in the `close()`? 


----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

For queries about this service, please contact Infrastructure at:
users@infra.apache.org

[GitHub] [pulsar] wolfstudy closed issue #7565: PulsarClient-cpp segv after close() method is called.

Posted by GitBox <gi...@apache.org>.

wolfstudy closed issue #7565:
URL: https://github.com/apache/pulsar/issues/7565


   


----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

For queries about this service, please contact Infrastructure at:
users@infra.apache.org

[GitHub] [pulsar] klwilson227 commented on issue #7565: PulsarClient-cpp segv after close() method is called.

Posted by GitBox <gi...@apache.org>.

klwilson227 commented on issue #7565:
URL: https://github.com/apache/pulsar/issues/7565#issuecomment-660407949


   With sleep in ClientConnection::Close() then After opening a consumer, and receiving messages. Close the consumer and client. 
   
   With the sleep in place the timers should pop and handlers waiting at mutex. When sleep releases the close completes and destroys the timers and then the handlers attempt to reset timers and the core dumps occur.
   
   
   
   Sent from my iPhone
   
   > On Jul 16, 2020, at 5:55 PM, Matteo Merli <no...@github.com> wrote:
   > 
   > 
   > @klwilson227 Do you have a simple way to reproduce this (even if it involves putting sleep() calls in the close()?
   > 
   > —
   > You are receiving this because you were mentioned.
   > Reply to this email directly, view it on GitHub, or unsubscribe.
   


----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

For queries about this service, please contact Infrastructure at:
users@infra.apache.org