Stefan, could you check if the attached patch mitigates the problem at least to some extend? Was suggested to me by Yann, all faults are mine. Thanks!
Cheers, Stefan
h2_session_clean.diff
Description: Binary data
> Am 19.01.2017 um 12:45 schrieb Stefan Priebe - Profihost AG > <s.pri...@profihost.ag>: > > Am 19.01.2017 um 11:56 schrieb Stefan Eissing: >> Stefan, >> >> yes, that is a known one that was addressed in v1.8.7. But I think it is >> related to the others. There is some mistake in my thinking about session >> pool cleanup that is more often uncovered by Yann's recent patches. I'll >> need some deep dive into this one. >> >> For you, that means v1.8.8 is the best right now. Hopefully we'll find this >> soon. > > But v1.8.8 is def. more often crashing than v1.8.3 which is included in > apache 2.4.25. > > With v.8.8 i see crashes like these which i don't have with 1.8.3: > (gdb) bt > #0 0x00007f6b5f9fef23 in apr_brigade_length () from > /usr/lib/x86_64-linux-gnu/libaprutil-1.so.0 > #1 0x000000000052afb6 in h2_util_bb_avail () > #2 0x00000000407e7a10 in ?? () > #3 0x00007f6b407e7a8c in ?? () > #4 0x00007f6b407e7a90 in ?? () > #5 0x00007f6b3e608668 in ?? () > #6 0x0000000000000000 in ?? () > > Greets, > Stefan > >> >> Cheers, >> >> Stefan >> >>> Am 19.01.2017 um 11:39 schrieb Stefan Priebe - Profihost AG >>> <s.pri...@profihost.ag>: >>> >>> Hi, >>> >>> with V7 and mod_http2 core i'm always seeing exactly THIS trace and >>> instead of every few minutes just once an hour: >>> >>> [Thread debugging using libthread_db enabled] >>> Using host libthread_db library "/lib/x86_64-linux-gnu/libthread_db.so.1". >>> Core was generated by `/usr/local/apache2/bin/httpd -k start'. >>> Program terminated with signal SIGSEGV, Segmentation fault. >>> #0 0x0000000000521d98 in h2_stream_out_prepare () >>> (gdb) bt >>> #0 0x0000000000521d98 in h2_stream_out_prepare () >>> #1 0x00007fb65bfeea80 in ?? () >>> #2 0x00007fb65bfeea8c in ?? () >>> #3 0x00007fb65bfeea90 in ?? () >>> #4 0x00007fb6599100a0 in ?? () >>> #5 0x00007fb659910f70 in ?? () >>> #6 0x00007fb65bfeeac0 in ?? () >>> #7 0x0000000000000000 in ?? () >>> >>> not all the others like with v1.8.8. So may this be a different one? >>> >>> Stefan >>> >>> Am 19.01.2017 um 09:21 schrieb Stefan Priebe - Profihost AG: >>>> And another one: >>>> >>>> #0 0x00007f7d81ba1f23 in apr_brigade_length () from >>>> /usr/lib/x86_64-linux-gnu/libaprutil-1.so.0 >>>> (gdb) bt >>>> #0 0x00007f7d81ba1f23 in apr_brigade_length () from >>>> /usr/lib/x86_64-linux-gnu/libaprutil-1.so.0 >>>> #1 0x000000000052a8bc in h2_util_bb_avail () >>>> #2 0x000000007112ca10 in ?? () >>>> #3 0x00007f7d7112ca8c in ?? () >>>> #4 0x00007f7d7112ca90 in ?? () >>>> #5 0x00007f7d60a4bad0 in ?? () >>>> #6 0x0000000000000000 in ?? () >>>> Am 19.01.2017 um 09:20 schrieb Stefan Priebe - Profihost AG: >>>>> Hi, >>>>> Am 19.01.2017 um 09:11 schrieb Yann Ylavic: >>>>>> On Thu, Jan 19, 2017 at 9:05 AM, Stefan Priebe - Profihost AG >>>>>> <s.pri...@profihost.ag> wrote: >>>>>>> With a vanilla apache 2.4.25 i got this one: >>>>>>> >>>>>>> Core was generated by `/usr/local/apache2/bin/httpd -k start'. >>>>>>> Program terminated with signal SIGSEGV, Segmentation fault. >>>>>>> #0 0x00007fe2bb7b8add in read () from >>>>>>> /lib/x86_64-linux-gnu/libpthread.so.0 >>>>>>> (gdb) bt >>>>>>> #0 0x00007fe2bb7b8add in read () from >>>>>>> /lib/x86_64-linux-gnu/libpthread.so.0 >>>>>>> #1 0x000000000048043a in ap_mpm_podx_check () >>>>>>> #2 <signal handler called> >>>>>>> #3 0x000000004b351bd0 in ?? () >>>>>>> Cannot access memory at address 0x0 >>>>>> >>>>>> This is probably not the faulting thread, could you find it with >>>>>> "thread apply all bt" or paste the whole output please? >>>>> >>>>> ah i found it via thread apply all bt: >>>>> >>>>> (gdb) bt >>>>> #0 0x0000000000521d98 in h2_stream_out_prepare () >>>>> #1 0x00007f5cbdfeaa80 in ?? () >>>>> #2 0x00007f5cbdfeaa8c in ?? () >>>>> #3 0x00007f5cbdfeaa90 in ?? () >>>>> #4 0x00007f5cb97ec0a0 in ?? () >>>>> #5 0x00007f5cb97ec988 in ?? () >>>>> #6 0x00007f5cbdfeaac0 in ?? () >>>>> #7 0x0000000000000000 in ?? () >>>>> >>>>> this is with a vanilla 2.4.25. >>>>> >>>>> Greets, >>>>> Stefan >>>>> >> >> Stefan Eissing >> >> <green/>bytes GmbH >> Hafenstrasse 16 >> 48155 Münster >> www.greenbytes.de >> Stefan Eissing <green/>bytes GmbH Hafenstrasse 16 48155 Münster www.greenbytes.de