--- Please enter the report below this line. ---
This crashes:
$ siege http://debian.home.lan/index.html
** SIEGE 2.70
** Preparing 15 concurrent users for battle.
The server is now under siege...Segmentation fault (core dumped)
This works:
$ siege http://debian/index.html
** SIEGE 2.70
** Preparing 15 concurrent users for battle.
The server is now under siege...
Both point to same IPv4 address:
$ host debian.home.lan
debian.home.lan has address 172.30.42.2
$ host debian
debian.home.lan has address 172.30.42.2
$ uname -a
Linux debian 3.6.6 #229 SMP PREEMPT Mon Nov 12 00:02:45 EET 2012 x86_64 GNU/Linux
.siegerc and /var/tmp/urls is attached.
Here is a stacktrace, unfortunately there is no -dbg package so no symbols:
(gdb) bt
#0 __memcpy_sse2 () at ../sysdeps/x86_64/multiarch/../memcpy.S:90
#1 0x000055555556695f in ?? ()
#2 0x000055555555d3cd in ?? ()
#3 0x000055555555dff5 in start_routine ()
#4 0x000055555555f1c5 in ?? ()
#5 0x00007ffff7bc7b50 in start_thread (arg=<optimized out>) at pthread_create.c:304
#6 0x00007ffff70cba7d in clone () at ../sysdeps/unix/sysv/linux/x86_64/clone.S:112
#7 0x0000000000000000 in ?? ()
If I compile siege 2.72 from upstream then there is no crash, so this seems to be a problem in Debian's 2.70 version.
Debian Release: wheezy/sid
500 unstable ftp.ro.debian.org
1 experimental ftp.ro.debian.org
--- Package information. ---
Depends (Version) | Installed
==========================-+-===========
libc6 (>= 2.7) |
libssl1.0.0 (>= 1.0.0) |
Package's Recommends field is empty.
Package's Suggests field is empty.
Hello, thanks for the detailed report. Right now, on a quick test with my default environment, I was not able to reproduce the bug, but I will spend more time on this tonight using all the details that you sent. Cheers,
... I am trying to reproduce with the latest versions in sid: $ dpkg -l libc6 ii libc6:amd64 2.13-37 amd64 $ dpkg -l libssl1.0.0 ii libssl1.0.0:am 1.0.1c-4 amd64 I have been trying, but I have not been able to reproduce the bug. Maybe you could try to reproduce it with upstream "2.70" version http://www.joedog.org/pub/siege/siege-2.70.tar.gz In the mean time I'll prepare a debug symbols package for you. Here is my output, using your ".siegerc" file: $ siege -R .siegerc http://www.jmaslibre.org/index.html ** SIEGE 2.70 ** Preparing 15 concurrent users for battle. The server is now under siege... Lifting the server siege... done. Transactions: 11 hits Availability: 100.00 % Elapsed time: 9.32 secs Data transferred: 0.00 MB Response time: 2.21 secs Transaction rate: 1.18 trans/sec Throughput: 0.00 MB/sec Concurrency: 2.61 Successful transactions: 11 Failed transactions: 0 Longest transaction: 5.37 Shortest transaction: 5.33 FILE: /var/tmp/siege.log You can disable this annoying message by editing the .siegerc file in your home directory; change the directive 'show-logfile' to false. $ siege -R .siegerc http://debian.home.lan/index.html ** SIEGE 2.70 ** Preparing 15 concurrent users for battle. The server is now under siege... Lifting the server siege... done. Transactions: 269 hits Availability: 100.00 % Elapsed time: 9.95 secs Data transferred: 0.04 MB Response time: 0.00 secs Transaction rate: 27.04 trans/sec Throughput: 0.00 MB/sec Concurrency: 0.11 Successful transactions: 269 Failed transactions: 0 Longest transaction: 0.02 Shortest transaction: 0.00 FILE: /var/tmp/siege.log You can disable this annoying message by editing the .siegerc file in your home directory; change the directive 'show-logfile' to false. Cheers,
I've got the same. It doesn't crash, and valgrind doesn't say anything either. I've done an 'apt-get source siege', and that doesn't crash either if I manually build it (using a simple ./configure && make). Maybe it has something to do with the way the Debian package is built (I see its using hardening-wrapper). I've rebuilt the Debian package using pbuilder in a wheezy chroot and a squeeze chroot. None of them crash. I've just checked again and /usr/bin/siege from the Debian repository still crashes. Do you remember what environment/compiler versions where used to build it? Because apparently simply rebuilding the package makes the crash go away, which makes me believe that it might be due to a compiler bug?