On 28.12.2017 14:53, Hilmar Preuße wrote: Добры бечер Олег,
On my system the addr2line command gives only useful results if I use the numbers in (), like this. At least output below makes sense to me. What does your addr2line returns? hille@amd64-sid:~$ addr2line -e /usr/sbin/proftpd 0x55e39fd633b0 ??:0 hille@amd64-sid:~$ addr2line -e /usr/sbin/proftpd 0x1f3b0 ./src/pool.c:546 Hilmar > Hi Олег, > > did to try to run addr2line as described on [1]? Does is print something > useful. The command is provided by package binutils. > > Hilmar > > [1] http://www.proftpd.org/docs/howto/Compiling.html > >> Got some positive results >> >> 2017-12-28 07:32:21,761 sim-ng-data proftpd[26614] sim-ng-data >> (PG10.local[192.168.3.17]): -----BEGIN STACK TRACE----- >> 2017-12-28 07:32:21,761 sim-ng-data proftpd[26614] sim-ng-data >> (PG10.local[192.168.3.17]): [0] /usr/sbin/proftpd(+0x1f3b0) >> [0x55e39fd633b0] >> 2017-12-28 07:32:21,761 sim-ng-data proftpd[26614] sim-ng-data >> (PG10.local[192.168.3.17]): [1] /usr/sbin/proftpd(+0x1f3b0) >> [0x55e39fd633b0] >> 2017-12-28 07:32:21,761 sim-ng-data proftpd[26614] sim-ng-data >> (PG10.local[192.168.3.17]): [2] /usr/sbin/proftpd(palloc+0x37) >> [0x55e39fd634b2] >> 2017-12-28 07:32:21,761 sim-ng-data proftpd[26614] sim-ng-data >> (PG10.local[192.168.3.17]): [3] /usr/sbin/proftpd(pcalloc+0x32) >> [0x55e39fd63547] >> 2017-12-28 07:32:21,761 sim-ng-data proftpd[26614] sim-ng-data >> (PG10.local[192.168.3.17]): [4] >> /usr/sbin/proftpd(pr_response_add+0xee) [0x55e39fd8de25] >> 2017-12-28 07:32:21,761 sim-ng-data proftpd[26614] sim-ng-data >> (PG10.local[192.168.3.17]): [5] /usr/lib/proftpd/mod_sftp.so(+0x3c7ef) >> [0x7f04249dd7ef] >> 2017-12-28 07:32:21,761 sim-ng-data proftpd[26614] sim-ng-data >> (PG10.local[192.168.3.17]): [6] >> /usr/lib/proftpd/mod_sftp.so(sftp_auth_handle+0x62) [0x7f04249ddc31] >> 2017-12-28 07:32:21,761 sim-ng-data proftpd[26614] sim-ng-data >> (PG10.local[192.168.3.17]): [7] >> /usr/lib/proftpd/mod_sftp.so(sftp_ssh2_packet_handle+0x363) >> [0x7f04249bc9ad] >> 2017-12-28 07:32:21,761 sim-ng-data proftpd[26614] sim-ng-data >> (PG10.local[192.168.3.17]): [8] /usr/lib/proftpd/mod_sftp.so(+0xfa07) >> [0x7f04249b0a07] >> 2017-12-28 07:32:21,761 sim-ng-data proftpd[26614] sim-ng-data >> (PG10.local[192.168.3.17]): [9] /usr/sbin/proftpd(+0x19ee3) >> [0x55e39fd5dee3] >> 2017-12-28 07:32:21,761 sim-ng-data proftpd[26614] sim-ng-data >> (PG10.local[192.168.3.17]): [10] /usr/sbin/proftpd(+0x1a706) >> [0x55e39fd5e706] >> 2017-12-28 07:32:21,761 sim-ng-data proftpd[26614] sim-ng-data >> (PG10.local[192.168.3.17]): [11] /usr/sbin/proftpd(+0x1bf2f) >> [0x55e39fd5ff2f] >> 2017-12-28 07:32:21,761 sim-ng-data proftpd[26614] sim-ng-data >> (PG10.local[192.168.3.17]): [12] /usr/sbin/proftpd(main+0x94b) >> [0x55e39fd60ce3] >> 2017-12-28 07:32:21,761 sim-ng-data proftpd[26614] sim-ng-data >> (PG10.local[192.168.3.17]): [13] >> /lib/x86_64-linux-gnu/libc.so.6(__libc_start_main+0xf1) [0x7f04282e1561] >> 2017-12-28 07:32:21,761 sim-ng-data proftpd[26614] sim-ng-data >> (PG10.local[192.168.3.17]): [14] /usr/sbin/proftpd(_start+0x2a) >> [0x55e39fd5adaa] >> 2017-12-28 07:32:21,761 sim-ng-data proftpd[26614] sim-ng-data >> (PG10.local[192.168.3.17]): -----END STACK TRACE----- >> 2017-12-28 07:32:21,761 sim-ng-data proftpd[26614] sim-ng-data >> (PG10.local[192.168.3.17]): ProFTPD terminating (signal 11) >> 2017-12-28 07:32:21,762 sim-ng-data proftpd[26614] sim-ng-data >> (PG10.local[192.168.3.17]): SSH2 session closed. >> >> >> 28.12.2017 0:09, Hilmar Preuße пишет: >>> On 26.12.2017 09:41, Мороз Олег wrote: >>> >>> Hi Олег, >>> >>>> Have made changes in >>>> >>>> /etc/security/limits.d/core_dump.conf >>>> >>>> then set path to save code dumps into /tmp >>>> >>>> vniiem@sim-ng-data /tmp % /sbin/sysctl -a -r kernel.core >>>> kernel.core_pattern = /tmp/core.%e.%p.%t >>>> kernel.core_pipe_limit = 0 >>>> kernel.core_uses_pid = 0 >>>> >>>> Then relogin and reinstall proftpd, tried to login to ftp , but no >>>> core.dump was generated >>>> >>> So, proftpd seems to trap the SIGSEGV (sig11) and performs a grace >>> shutdown upon SIG11. I've found [1] giving some hints to debug crash >>> problems. I've put debug binaries using --enable-devel=stacktrace on >>> [2]. I've crashed the binaries using a SIGSEGV and found a stack trace >>> in /var/log/proftpd/proftpd.log: >>> >>> 2017-12-27 21:51:07,528 amd64-sid proftpd[23112] amd64-sid: -----BEGIN >>> STACK TRACE----- >>> 2017-12-27 21:51:07,528 amd64-sid proftpd[23112] amd64-sid: [0] >>> /lib/x86_64-linux-gnu/libc.so.6(__select+0x13) [0x7fd99e1c3cc3] >>> 2017-12-27 21:51:07,529 amd64-sid proftpd[23112] amd64-sid: [1] >>> /lib/x86_64-linux-gnu/libc.so.6(__select+0x13) [0x7fd99e1c3cc3] >>> 2017-12-27 21:51:07,529 amd64-sid proftpd[23112] amd64-sid: [2] >>> /usr/sbin/proftpd(+0x1a2d0) [0x5559bc50c2d0] >>> 2017-12-27 21:51:07,529 amd64-sid proftpd[23112] amd64-sid: [3] >>> /usr/sbin/proftpd(+0x1bf2f) [0x5559bc50df2f] >>> 2017-12-27 21:51:07,529 amd64-sid proftpd[23112] amd64-sid: [4] >>> /usr/sbin/proftpd(main+0x94b) [0x5559bc50ece3] >>> 2017-12-27 21:51:07,529 amd64-sid proftpd[23112] amd64-sid: [5] >>> /lib/x86_64-linux-gnu/libc.so.6(__libc_start_main+0xf1) [0x7fd99e0ff561] >>> 2017-12-27 21:51:07,529 amd64-sid proftpd[23112] amd64-sid: [6] >>> /usr/sbin/proftpd(_start+0x2a) [0x5559bc508daa] >>> 2017-12-27 21:51:07,529 amd64-sid proftpd[23112] amd64-sid: -----END >>> STACK TRACE----- >>> >>> Please install the binary, the stack trace should at least show us if >>> the problematic code is in proftp or anywhere else. >>> The addr2line directions in [1] did not help me, I always got "??:0". At >>> least the stack trace should be a basis to issue a bug @upstream. >>> >>> Many thanks for help! >>> >>> Hilmar >>> >>> [1] http://www.proftpd.org/docs/howto/Compiling.html >>> [2] https://freeshell.de/~hille42/proftpd/stacktrace/ >> > > -- #206401 http://counter.li.org