Bug #412972 “bad signal mask of ssh sessions” : Bugs : openssh package : Ubuntu

Revision history for this message

In OpenSSH Portable Bugzilla #271, Damien Miller (djm) wrote on 2003-01-07:

#1

We already reset SIGCHLD to SIG_DFL in main(), maybe we don't do it
early enough...

Revision history for this message

In OpenSSH Portable Bugzilla #271, Damien Miller (djm) wrote on 2003-05-16:

#2

Could you verify this with a recent version, preferably including diag
output.

Revision history for this message

In OpenSSH Portable Bugzilla #271, Damien Miller (djm) wrote on 2005-02-09:

#3

20 months + no reply == closed bug

Revision history for this message

In OpenSSH Portable Bugzilla #271, Darren Tucker (dtucker) wrote on 2006-10-07:

#4

Change all RESOLVED bug to CLOSED with the exception of the ones fixed
post-4.4.

Revision history for this message

Michael Helmling (supermihi) wrote on 2009-08-13: can only kill processes with -9 in karmic

#5

On one of my systems that runs the karmic alpha already, the "kill" command has no effect unless used with the -9 option. I assume this might be related to https://bugs.launchpad.net/ubuntu/+source/openssh/+bug/407428

My steps to reproduce on this machine: "sleep 100" on one terminal, "killall sleep" on another. System is up to date, running kernel 2.6.31-4-generic. Sorry for having no idea which package this corresponds to!

Revision history for this message

Jonathan Marsden (jmarsden) wrote on 2009-08-17:

#6

I cannot reproduce this here.

sleep 100 &
killall sleep

works fine for me in Ubuntu Karmic Alpha 4 amd64 (with updates as of 2009-08-16):

  Script started on Sun 16 Aug 2009 06:54:43 PM PDT
  jonathan@black:~$ sleep 100 &
  [1] 28262
  jonathan@black:~$ killall sleep
  [1]+ Terminated sleep 100
  jonathan@black:~$ exit
  exit

Script done on Sun 16 Aug 2009 06:55:00 PM PDT

Doing it with the sleep 100 in a separate terminal also works fine for me.

Revision history for this message

Jonathan Marsden (jmarsden) wrote on 2009-08-17:

#7

In case it makes a difference, I was using a virtualbox VM rather than a physical machine to run Karmic Alpha 4 for the above test.

Revision history for this message

Victor Vargas (kamus) wrote on 2009-08-18:

#8

Same here, I try to reproduce this in Karmic Alpha up to date (running under VirtualBox) and works fine for me too.

affects:

ubuntu → procps (Ubuntu)

Revision history for this message

Michael Helmling (supermihi) wrote on 2009-08-18:

#9

Strange .. could you give me any hints how I could examine further why this happens?

Revision history for this message

Jonathan Marsden (jmarsden) wrote on 2009-08-19:

#10

Michael wrote: "could you give me any hints how I could examine further why this happens?"

The first thing might be to see if it is replicable -- can you install Karmic Alpha 4 on another machine and check whether you get the same issue? If not, then either it relates to the current install you have, or to the hardware you installed it on. If you can duplicate it, document *exactly* how you are doing the install process, what you are choosing at every step of the way, so that others can follow your documented steps and reproduce this.

Until someone other than you can duplicate the issue, it is unlikely we'll be able to provide an explanation (or even hints) as to why it is happening. So far, it has happened exactly one on one install on one piece of hardware... we need to determine if that was "just a one off", or if there is a reproduceable issue here.

Revision history for this message

Matti Hiljanen (matti-hiljanen) wrote on 2009-08-20:

#11

I'm seeing this too, clean up-to-date install of karmic alpha4 amd64 (netinst with mini.iso) on a physical machine.

This happens only when logged in via ssh, doing the test locally (over a serial console) works. Definitely seems to be related to https://bugs.launchpad.net/ubuntu/+source/openssh/+bug/407428

Jonathan Marsden (jmarsden) on 2009-08-21

summary:

- can only kill processes with -9 in karmic
+ can only kill processes with -9 in karmic from SSH sessions

Revision history for this message

Scott James Remnant (Canonical) (canonical-scott) wrote on 2009-08-25: Re: can only kill processes with -9 in karmic from SSH sessions

#12

Pick on a process to kill, then run:

strace kill -TERM pid-of-process

And attach the output from strace

Changed in procps (Ubuntu):
importance:	Undecided → Medium
status:	New → Incomplete

Revision history for this message

Michael Helmling (supermihi) wrote on 2009-08-25:

#13

Download full text (4.2 KiB)

Hi,
here's what I get (23473 is a "sleep 100" process):

helmling@menk:~$ strace kill -TERM 23473
execve("/bin/kill", ["kill", "-TERM", "23473"], [/* 18 vars */]) = 0
brk(0) = 0x762000
mmap(NULL, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) = 0x7f6fff412000
access("/etc/ld.so.nohwcap", F_OK) = -1 ENOENT (No such file or directory)
mmap(NULL, 8192, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) = 0x7f6fff410000
access("/etc/ld.so.preload", R_OK) = -1 ENOENT (No such file or directory)
open("/etc/ld.so.cache", O_RDONLY) = 3
fstat(3, {st_mode=S_IFREG|0644, st_size=125575, ...}) = 0
mmap(NULL, 125575, PROT_READ, MAP_PRIVATE, 3, 0) = 0x7f6fff3f1000
close(3) = 0
access("/etc/ld.so.nohwcap", F_OK) = -1 ENOENT (No such file or directory)
open("/lib/libproc-3.2.8.so", O_RDONLY) = 3
read(3, "\177ELF\2\1\1\0\0\0\0\0\0\0\0\0\3\0>\0\1\0\0\0\240A\0\0\0\0\0\0"..., 832) = 832
fstat(3, {st_mode=S_IFREG|0644, st_size=76696, ...}) = 0
mmap(NULL, 2249496, PROT_READ|PROT_EXEC, MAP_PRIVATE|MAP_DENYWRITE, 3, 0) = 0x7f6ffefcf000
mprotect(0x7f6ffefe0000, 2097152, PROT_NONE) = 0
mmap(0x7f6fff1e0000, 8192, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_FIXED|MAP_DENYWRITE, 3, 0x11000) = 0x7f6fff1e0000
mmap(0x7f6fff1e2000, 74520, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_FIXED|MAP_ANONYMOUS, -1, 0) = 0x7f6fff1e2000
close(3) = 0
access("/etc/ld.so.nohwcap", F_OK) = -1 ENOENT (No such file or directory)
open("/lib/libc.so.6", O_RDONLY) = 3
read(3, "\177ELF\2\1\1\0\0\0\0\0\0\0\0\0\3\0>\0\1\0\0\0\340\353\1\0\0\0\0\0"..., 832) = 832
fstat(3, {st_mode=S_IFREG|0755, st_size=1490312, ...}) = 0
mmap(NULL, 3598344, PROT_READ|PROT_EXEC, MAP_PRIVATE|MAP_DENYWRITE, 3, 0) = 0x7f6ffec60000
mprotect(0x7f6ffedc6000, 2093056, PROT_NONE) = 0
mmap(0x7f6ffefc5000, 20480, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_FIXED|MAP_DENYWRITE, 3, 0x165000) = 0x7f6ffefc5000
mmap(0x7f6ffefca000, 18440, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_FIXED|MAP_ANONYMOUS, -1, 0) = 0x7f6ffefca000
close(3) = 0
mmap(NULL, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) = 0x7f6fff3f0000
mmap(NULL, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) = 0x7f6fff3ef000
arch_prctl(ARCH_SET_FS, 0x7f6fff3ef6f0) = 0
mprotect(0x7f6ffefc5000, 16384, PROT_READ) = 0
mprotect(0x7f6fff1e0000, 4096, PROT_READ) = 0
mprotect(0x603000, 4096, PROT_READ) = 0
mprotect(0x7f6fff413000, 4096, PROT_READ) = 0
munmap(0x7f6fff3f1000, 125575) = 0
brk(0) = 0x762000
brk(0x783000) = 0x783000
open("/proc...

Hi,
here's what I get (23473 is a "sleep 100" process):

helmling@menk:~$ strace kill -TERM 23473
execve("/bin/kill", ["kill", "-TERM", "23473"], [/* 18 vars */]) = 0
brk(0)                                  = 0x762000                  
mmap(NULL, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) = 0x7f6fff412000
access("/etc/ld.so.nohwcap", F_OK)      = -1 ENOENT (No such file or directory)          
mmap(NULL, 8192, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) = 0x7f6fff410000
access("/etc/ld.so.preload", R_OK)      = -1 ENOENT (No such file or directory)          
open("/etc/ld.so.cache", O_RDONLY)      = 3                                              
fstat(3, {st_mode=S_IFREG|0644, st_size=125575, ...}) = 0                                
mmap(NULL, 125575, PROT_READ, MAP_PRIVATE, 3, 0) = 0x7f6fff3f1000                        
close(3)                                = 0                                              
access("/etc/ld.so.nohwcap", F_OK)      = -1 ENOENT (No such file or directory)          
open("/lib/libproc-3.2.8.so", O_RDONLY) = 3                                              
read(3, "\177ELF\2\1\1\0\0\0\0\0\0\0\0\0\3\0>\0\1\0\0\0\240A\0\0\0\0\0\0"..., 832) = 832 
fstat(3, {st_mode=S_IFREG|0644, st_size=76696, ...}) = 0                                 
mmap(NULL, 2249496, PROT_READ|PROT_EXEC, MAP_PRIVATE|MAP_DENYWRITE, 3, 0) = 0x7f6ffefcf000
mprotect(0x7f6ffefe0000, 2097152, PROT_NONE) = 0                                          
mmap(0x7f6fff1e0000, 8192, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_FIXED|MAP_DENYWRITE, 3, 0x11000) = 0x7f6fff1e0000
mmap(0x7f6fff1e2000, 74520, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_FIXED|MAP_ANONYMOUS, -1, 0) = 0x7f6fff1e2000    
close(3)                                = 0                                                                       
access("/etc/ld.so.nohwcap", F_OK)      = -1 ENOENT (No such file or directory)                                   
open("/lib/libc.so.6", O_RDONLY)        = 3                                                                       
read(3, "\177ELF\2\1\1\0\0\0\0\0\0\0\0\0\3\0>\0\1\0\0\0\340\353\1\0\0\0\0\0"..., 832) = 832
fstat(3, {st_mode=S_IFREG|0755, st_size=1490312, ...}) = 0
mmap(NULL, 3598344, PROT_READ|PROT_EXEC, MAP_PRIVATE|MAP_DENYWRITE, 3, 0) = 0x7f6ffec60000
mprotect(0x7f6ffedc6000, 2093056, PROT_NONE) = 0
mmap(0x7f6ffefc5000, 20480, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_FIXED|MAP_DENYWRITE, 3, 0x165000) = 0x7f6ffefc5000
mmap(0x7f6ffefca000, 18440, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_FIXED|MAP_ANONYMOUS, -1, 0) = 0x7f6ffefca000
close(3)                                = 0
mmap(NULL, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) = 0x7f6fff3f0000
mmap(NULL, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) = 0x7f6fff3ef000
arch_prctl(ARCH_SET_FS, 0x7f6fff3ef6f0) = 0
mprotect(0x7f6ffefc5000, 16384, PROT_READ) = 0
mprotect(0x7f6fff1e0000, 4096, PROT_READ) = 0
mprotect(0x603000, 4096, PROT_READ)     = 0
mprotect(0x7f6fff413000, 4096, PROT_READ) = 0
munmap(0x7f6fff3f1000, 125575)          = 0
brk(0)                                  = 0x762000
brk(0x783000)                           = 0x783000
open("/proc/version", O_RDONLY)         = 3
fstat(3, {st_mode=S_IFREG|0444, st_size=0, ...}) = 0
mmap(NULL, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) = 0x7f6fff40f000
read(3, "Linux version 2.6.31-6-generic ("..., 1024) = 136
close(3)                                = 0
munmap(0x7f6fff40f000, 4096)            = 0
open("/proc/stat", O_RDONLY|O_CLOEXEC)  = 3
read(3, "cpu  258654 64949 198562 1417529"..., 8192) = 1188
close(3)                                = 0
getpid()                                = 32478
kill(23473, SIGTERM)                    = -1 ESRCH (No such process)
dup(2)                                  = 3
fcntl(3, F_GETFL)                       = 0x8002 (flags O_RDWR|O_LARGEFILE)
fstat(3, {st_mode=S_IFCHR|0620, st_rdev=makedev(136, 1), ...}) = 0
mmap(NULL, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) = 0x7f6fff40f000
lseek(3, 0, SEEK_CUR)                   = -1 ESPIPE (Illegal seek)
write(3, "kill: No such process\n", 22kill: No such process
) = 22
close(3)                                = 0
munmap(0x7f6fff40f000, 4096)            = 0
exit_group(1)                           = ?

Revision history for this message

Scott James Remnant (Canonical) (canonical-scott) wrote on 2009-08-25:

#14

The key line here is:

kill(23473, SIGTERM) = -1 ESRCH (No such process)

Could you do the same again, this time run both "strace kill -TERM pid" and then immediately after "strace kill -9 pid" to prove that -9 works

summary:

- can only kill processes with -9 in karmic from SSH sessions
+ can only kill processes with -9 in karmic from SSH sessions, -TERM does
+ not work

Revision history for this message

Michael Helmling (supermihi) wrote on 2009-08-26: Re: can only kill processes with -9 in karmic from SSH sessions, -TERM does not work

#15

Download full text (9.7 KiB)

Here you go...

root@menk:~# ps -A|grep sleep
3644 pts/0 00:00:00 sleep

root@menk:~# strace kill -TERM 3644
execve("/bin/kill", ["kill", "-TERM", "3644"], [/* 18 vars */]) = 0
brk(0) = 0x8ee000
mmap(NULL, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) = 0x7ff12fa8d000
access("/etc/ld.so.nohwcap", F_OK) = -1 ENOENT (No such file or directory)
mmap(NULL, 8192, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) = 0x7ff12fa8b000
access("/etc/ld.so.preload", R_OK) = -1 ENOENT (No such file or directory)
open("/etc/ld.so.cache", O_RDONLY) = 3
fstat(3, {st_mode=S_IFREG|0644, st_size=124031, ...}) = 0
mmap(NULL, 124031, PROT_READ, MAP_PRIVATE, 3, 0) = 0x7ff12fa6c000
close(3) = 0
access("/etc/ld.so.nohwcap", F_OK) = -1 ENOENT (No such file or directory)
open("/lib/libproc-3.2.8.so", O_RDONLY) = 3
read(3, "\177ELF\2\1\1\0\0\0\0\0\0\0\0\0\3\0>\0\1\0\0\0\240A\0\0\0\0\0\0"..., 832) = 832
fstat(3, {st_mode=S_IFREG|0644, st_size=76696, ...}) = 0
mmap(NULL, 2249496, PROT_READ|PROT_EXEC, MAP_PRIVATE|MAP_DENYWRITE, 3, 0) = 0x7ff12f64a000
mprotect(0x7ff12f65b000, 2097152, PROT_NONE) = 0
mmap(0x7ff12f85b000, 8192, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_FIXED|MAP_DENYWRITE, 3, 0x11000) = 0x7ff12f85b000
mmap(0x7ff12f85d000, 74520, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_FIXED|MAP_ANONYMOUS, -1, 0) = 0x7ff12f85d000
close(3) = 0
access("/etc/ld.so.nohwcap", F_OK) = -1 ENOENT (No such file or directory)
open("/lib/libc.so.6", O_RDONLY) = 3
read(3, "\177ELF\2\1\1\0\0\0\0\0\0\0\0\0\3\0>\0\1\0\0\0\340\353\1\0\0\0\0\0"..., 832) = 832
fstat(3, {st_mode=S_IFREG|0755, st_size=1490312, ...}) = 0
mmap(NULL, 3598344, PROT_READ|PROT_EXEC, MAP_PRIVATE|MAP_DENYWRITE, 3, 0) = 0x7ff12f2db000
mprotect(0x7ff12f441000, 2093056, PROT_NONE) = 0
mmap(0x7ff12f640000, 20480, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_FIXED|MAP_DENYWRITE, 3, 0x165000) = 0x7ff12f640000
mmap(0x7ff12f645000, 18440, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_FIXED|MAP_ANONYMOUS, -1, 0) = 0x7ff12f645000
close(3) = 0
mmap(NULL, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) = 0x7ff12fa6b000
mmap(NULL, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) = 0x7ff12fa6a000
arch_prctl(ARCH_SET_FS, 0x7ff12fa6a6f0) = 0 ...

Here you go...

root@menk:~# ps -A|grep sleep
 3644 pts/0    00:00:00 sleep

root@menk:~# strace kill -TERM 3644
execve("/bin/kill", ["kill", "-TERM", "3644"], [/* 18 vars */]) = 0
brk(0)                                  = 0x8ee000                 
mmap(NULL, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) = 0x7ff12fa8d000
access("/etc/ld.so.nohwcap", F_OK)      = -1 ENOENT (No such file or directory)          
mmap(NULL, 8192, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) = 0x7ff12fa8b000
access("/etc/ld.so.preload", R_OK)      = -1 ENOENT (No such file or directory)          
open("/etc/ld.so.cache", O_RDONLY)      = 3                                              
fstat(3, {st_mode=S_IFREG|0644, st_size=124031, ...}) = 0                                
mmap(NULL, 124031, PROT_READ, MAP_PRIVATE, 3, 0) = 0x7ff12fa6c000                        
close(3)                                = 0                                              
access("/etc/ld.so.nohwcap", F_OK)      = -1 ENOENT (No such file or directory)          
open("/lib/libproc-3.2.8.so", O_RDONLY) = 3                                              
read(3, "\177ELF\2\1\1\0\0\0\0\0\0\0\0\0\3\0>\0\1\0\0\0\240A\0\0\0\0\0\0"..., 832) = 832 
fstat(3, {st_mode=S_IFREG|0644, st_size=76696, ...}) = 0                                 
mmap(NULL, 2249496, PROT_READ|PROT_EXEC, MAP_PRIVATE|MAP_DENYWRITE, 3, 0) = 0x7ff12f64a000
mprotect(0x7ff12f65b000, 2097152, PROT_NONE) = 0                                          
mmap(0x7ff12f85b000, 8192, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_FIXED|MAP_DENYWRITE, 3, 0x11000) = 0x7ff12f85b000
mmap(0x7ff12f85d000, 74520, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_FIXED|MAP_ANONYMOUS, -1, 0) = 0x7ff12f85d000    
close(3)                                = 0                                                                       
access("/etc/ld.so.nohwcap", F_OK)      = -1 ENOENT (No such file or directory)                                   
open("/lib/libc.so.6", O_RDONLY)        = 3                                                                       
read(3, "\177ELF\2\1\1\0\0\0\0\0\0\0\0\0\3\0>\0\1\0\0\0\340\353\1\0\0\0\0\0"..., 832) = 832                       
fstat(3, {st_mode=S_IFREG|0755, st_size=1490312, ...}) = 0                                                        
mmap(NULL, 3598344, PROT_READ|PROT_EXEC, MAP_PRIVATE|MAP_DENYWRITE, 3, 0) = 0x7ff12f2db000                        
mprotect(0x7ff12f441000, 2093056, PROT_NONE) = 0                                                                  
mmap(0x7ff12f640000, 20480, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_FIXED|MAP_DENYWRITE, 3, 0x165000) = 0x7ff12f640000
mmap(0x7ff12f645000, 18440, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_FIXED|MAP_ANONYMOUS, -1, 0) = 0x7ff12f645000      
close(3)                                = 0                                                                         
mmap(NULL, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) = 0x7ff12fa6b000                           
mmap(NULL, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) = 0x7ff12fa6a000                           
arch_prctl(ARCH_SET_FS, 0x7ff12fa6a6f0) = 0                                                                         
mprotect(0x7ff12f640000, 16384, PROT_READ) = 0                                                                      
mprotect(0x7ff12f85b000, 4096, PROT_READ) = 0                                                                       
mprotect(0x603000, 4096, PROT_READ)     = 0                                                                         
mprotect(0x7ff12fa8e000, 4096, PROT_READ) = 0                                                                       
munmap(0x7ff12fa6c000, 124031)          = 0                                                                         
brk(0)                                  = 0x8ee000                                                                  
brk(0x90f000)                           = 0x90f000                                                                  
open("/proc/version", O_RDONLY)         = 3                                                                         
fstat(3, {st_mode=S_IFREG|0444, st_size=0, ...}) = 0                                                                
mmap(NULL, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) = 0x7ff12fa8a000                           
read(3, "Linux version 2.6.31-6-generic ("..., 1024) = 136                                                          
close(3)                                = 0                                                                         
munmap(0x7ff12fa8a000, 4096)            = 0                                                                         
open("/proc/stat", O_RDONLY|O_CLOEXEC)  = 3                                                                         
read(3, "cpu  444 0 603 18519 4676 9 15 0"..., 8192) = 1096                                                         
close(3)                                = 0                                                                         
getpid()                                = 3648                                                                      
kill(3644, SIGTERM)                     = 0                                                                         
exit_group(0)                           = ?

root@menk:~# strace kill -9 3644                                                                                
execve("/bin/kill", ["kill", "-9", "3644"], [/* 18 vars */]) = 0                                                
brk(0)                                  = 0x1488000                                                             
mmap(NULL, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) = 0x7f6616ec7000                       
access("/etc/ld.so.nohwcap", F_OK)      = -1 ENOENT (No such file or directory)                                 
mmap(NULL, 8192, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) = 0x7f6616ec5000                       
access("/etc/ld.so.preload", R_OK)      = -1 ENOENT (No such file or directory)                                 
open("/etc/ld.so.cache", O_RDONLY)      = 3                                                                     
fstat(3, {st_mode=S_IFREG|0644, st_size=124031, ...}) = 0                                                       
mmap(NULL, 124031, PROT_READ, MAP_PRIVATE, 3, 0) = 0x7f6616ea6000                                               
close(3)                                = 0                                                                     
access("/etc/ld.so.nohwcap", F_OK)      = -1 ENOENT (No such file or directory)                                 
open("/lib/libproc-3.2.8.so", O_RDONLY) = 3                                                                     
read(3, "\177ELF\2\1\1\0\0\0\0\0\0\0\0\0\3\0>\0\1\0\0\0\240A\0\0\0\0\0\0"..., 832) = 832                        
fstat(3, {st_mode=S_IFREG|0644, st_size=76696, ...}) = 0                                                        
mmap(NULL, 2249496, PROT_READ|PROT_EXEC, MAP_PRIVATE|MAP_DENYWRITE, 3, 0) = 0x7f6616a84000                      
mprotect(0x7f6616a95000, 2097152, PROT_NONE) = 0                                                                
mmap(0x7f6616c95000, 8192, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_FIXED|MAP_DENYWRITE, 3, 0x11000) = 0x7f6616c95000
mmap(0x7f6616c97000, 74520, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_FIXED|MAP_ANONYMOUS, -1, 0) = 0x7f6616c97000    
close(3)                                = 0                                                                       
access("/etc/ld.so.nohwcap", F_OK)      = -1 ENOENT (No such file or directory)                                   
open("/lib/libc.so.6", O_RDONLY)        = 3                                                                       
read(3, "\177ELF\2\1\1\0\0\0\0\0\0\0\0\0\3\0>\0\1\0\0\0\340\353\1\0\0\0\0\0"..., 832) = 832                       
fstat(3, {st_mode=S_IFREG|0755, st_size=1490312, ...}) = 0                                                        
mmap(NULL, 3598344, PROT_READ|PROT_EXEC, MAP_PRIVATE|MAP_DENYWRITE, 3, 0) = 0x7f6616715000                        
mprotect(0x7f661687b000, 2093056, PROT_NONE) = 0                                                                  
mmap(0x7f6616a7a000, 20480, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_FIXED|MAP_DENYWRITE, 3, 0x165000) = 0x7f6616a7a000
mmap(0x7f6616a7f000, 18440, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_FIXED|MAP_ANONYMOUS, -1, 0) = 0x7f6616a7f000      
close(3)                                = 0                                                                         
mmap(NULL, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) = 0x7f6616ea5000
mmap(NULL, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) = 0x7f6616ea4000
arch_prctl(ARCH_SET_FS, 0x7f6616ea46f0) = 0
mprotect(0x7f6616a7a000, 16384, PROT_READ) = 0
mprotect(0x7f6616c95000, 4096, PROT_READ) = 0
mprotect(0x603000, 4096, PROT_READ)     = 0
mprotect(0x7f6616ec8000, 4096, PROT_READ) = 0
munmap(0x7f6616ea6000, 124031)          = 0
brk(0)                                  = 0x1488000
brk(0x14a9000)                          = 0x14a9000
open("/proc/version", O_RDONLY)         = 3
fstat(3, {st_mode=S_IFREG|0444, st_size=0, ...}) = 0
mmap(NULL, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) = 0x7f6616ec4000
read(3, "Linux version 2.6.31-6-generic ("..., 1024) = 136
close(3)                                = 0
munmap(0x7f6616ec4000, 4096)            = 0
open("/proc/stat", O_RDONLY|O_CLOEXEC)  = 3
read(3, "cpu  446 0 604 22290 4676 10 15 "..., 8192) = 1098
close(3)                                = 0
getpid()                                = 3650
kill(3644, SIGKILL)                     = 0
exit_group(0)                           = ?

Revision history for this message

Michael Helmling (supermihi) wrote on 2009-08-26:

#16

Actually the "no such process" line is not present anymore here, but the symptoms are the same -- sleep runs until I kill with -9.

Revision history for this message

Scott James Remnant (Canonical) (canonical-scott) wrote on 2009-08-27: Re: [Bug 412972] Re: can only kill processes with -9 in karmic from SSH sessions, -TERM does not work

#17

On Wed, 2009-08-26 at 08:22 +0000, Michael Helmling wrote:

> Actually the "no such process" line is not present anymore here, but the
> symptoms are the same -- sleep runs until I kill with -9.
>
Right, I suspect that you just weren't quick enough that time ;-)

Could you confirm a few things for me:

- what kind of machine is this? PC, Mac, emulated? virtual machine?
etc.

- how are you running the "sleep" and "kill" commands? from X? from a
getty? from ssh?

Could you run a "sleep" as before, and this time "cat /proc/PID/status"
from another terminal and attach that.

Thanks,

Scott
--
Scott James Remnant
<email address hidden>

Revision history for this message

Michael Helmling (supermihi) wrote on 2009-08-27: Re: can only kill processes with -9 in karmic from SSH sessions, -TERM does not work

#18

Hi,
the machine is a PC (amd64), not virtual. All I said before was regarding SSH sessions only, but right now I happen to be at the site, so I'm able to do some further testing. The following holds:

sleep on SSH, kill on SSH: not working
sleep on SSH, kill on getty: not working
sleep on getty, kill on SSH: working
sleep on getty, kill on getty: working

So this is somewhat SSH related, and I even more suspect a relation to bug #407428. However, the mentioned bug is gone after the last update+reboot, but this one is NOT.

Here is your /proc/status from a sleep running at SSH:

# cat /proc/4781/status
Name: sleep
State: S (sleeping)
Tgid: 4781
Pid: 4781
PPid: 3695
TracerPid: 0
Uid: 0 0 0 0
Gid: 0 0 0 0
FDSize: 256
Groups: 0
VmPeak: 9436 kB
VmSize: 9436 kB
VmLck: 0 kB
VmHWM: 804 kB
VmRSS: 804 kB
VmData: 192 kB
VmStk: 84 kB
VmExe: 32 kB
VmLib: 1676 kB
VmPTE: 40 kB
Threads: 1
SigQ: 7/16382
SigPnd: 0000000000000000
ShdPnd: 0000000000000000
SigBlk: fffffffe7ffadeff
SigIgn: 0000000000000000
SigCgt: 0000000180000000
CapInh: 0000000000000000
CapPrm: ffffffffffffffff
CapEff: ffffffffffffffff
CapBnd: ffffffffffffffff
Cpus_allowed: 3
Cpus_allowed_list: 0-1
Mems_allowed: 00000000,00000001
Mems_allowed_list: 0
voluntary_ctxt_switches: 2
nonvoluntary_ctxt_switches: 2

Revision history for this message

Colin Watson (cjwatson) wrote on 2009-08-27:

#19

Might any of you have restarted sshd from within a su session? I don't have definite proof that it's related, but I note that the quoted blocked signal mask corresponds exactly to those signals that su blocks. I wonder if it's due to some PAM session module ...

Revision history for this message

Scott James Remnant (Canonical) (canonical-scott) wrote on 2009-08-27:

#20

Michael: could you also give us the /proc/PID/status of a sleep after you've sent it the TERM signal

summary:

- can only kill processes with -9 in karmic from SSH sessions, -TERM does
- not work
+ bad signal mask of ssh sessions

Revision history for this message

Michael Helmling (supermihi) wrote on 2009-09-01:

#21

Here you go ...

# cat /proc/12218/status
Name: sleep
State: S (sleeping)
Tgid: 12218
Pid: 12218
PPid: 12202
TracerPid: 0
Uid: 0 0 0 0
Gid: 0 0 0 0
FDSize: 256
Groups: 0
VmPeak: 9436 kB
VmSize: 9436 kB
VmLck: 0 kB
VmHWM: 804 kB
VmRSS: 804 kB
VmData: 192 kB
VmStk: 84 kB
VmExe: 32 kB
VmLib: 1676 kB
VmPTE: 40 kB
Threads: 1
SigQ: 10/16382
SigPnd: 0000000000000000
ShdPnd: 0000000000000000
SigBlk: fffffffe7ffadeff
SigIgn: 0000000000000000
SigCgt: 0000000180000000
CapInh: 0000000000000000
CapPrm: ffffffffffffffff
CapEff: ffffffffffffffff
CapBnd: ffffffffffffffff
Cpus_allowed: 3
Cpus_allowed_list: 0-1
Mems_allowed: 00000000,00000001
Mems_allowed_list: 0
voluntary_ctxt_switches: 1
nonvoluntary_ctxt_switches: 1
root@menk:~# ps -A|grep sleep
12218 pts/3 00:00:00 sleep
root@menk:~# kill 12218
root@menk:~# cat /proc/12218/status
Name: sleep
State: S (sleeping)
Tgid: 12218
Pid: 12218
PPid: 12202
TracerPid: 0
Uid: 0 0 0 0
Gid: 0 0 0 0
FDSize: 256
Groups: 0
VmPeak: 9436 kB
VmSize: 9436 kB
VmLck: 0 kB
VmHWM: 804 kB
VmRSS: 804 kB
VmData: 192 kB
VmStk: 84 kB
VmExe: 32 kB
VmLib: 1676 kB
VmPTE: 40 kB
Threads: 1
SigQ: 11/16382
SigPnd: 0000000000000000
ShdPnd: 0000000000004000
SigBlk: fffffffe7ffadeff
SigIgn: 0000000000000000
SigCgt: 0000000180000000
CapInh: 0000000000000000
CapPrm: ffffffffffffffff
CapEff: ffffffffffffffff
CapBnd: ffffffffffffffff
Cpus_allowed: 3
Cpus_allowed_list: 0-1
Mems_allowed: 00000000,00000001
Mems_allowed_list: 0
voluntary_ctxt_switches: 1
nonvoluntary_ctxt_switches: 1

Here you go ...

# cat /proc/12218/status 
Name:   sleep                       
State:  S (sleeping)                
Tgid:   12218                       
Pid:    12218                       
PPid:   12202                       
TracerPid:      0                   
Uid:    0       0       0       0   
Gid:    0       0       0       0   
FDSize: 256                         
Groups: 0                           
VmPeak:     9436 kB                 
VmSize:     9436 kB                 
VmLck:         0 kB                 
VmHWM:       804 kB                 
VmRSS:       804 kB                 
VmData:      192 kB                 
VmStk:        84 kB                 
VmExe:        32 kB                 
VmLib:      1676 kB                 
VmPTE:        40 kB                 
Threads:        1                   
SigQ:   10/16382                    
SigPnd: 0000000000000000            
ShdPnd: 0000000000000000            
SigBlk: fffffffe7ffadeff            
SigIgn: 0000000000000000            
SigCgt: 0000000180000000            
CapInh: 0000000000000000            
CapPrm: ffffffffffffffff            
CapEff: ffffffffffffffff            
CapBnd: ffffffffffffffff            
Cpus_allowed:   3                   
Cpus_allowed_list:      0-1         
Mems_allowed:   00000000,00000001   
Mems_allowed_list:      0           
voluntary_ctxt_switches:        1   
nonvoluntary_ctxt_switches:     1   
root@menk:~# ps -A|grep sleep
12218 pts/3    00:00:00 sleep
root@menk:~# kill 12218
root@menk:~# cat /proc/12218/status 
Name:   sleep                       
State:  S (sleeping)                
Tgid:   12218                       
Pid:    12218                       
PPid:   12202                       
TracerPid:      0                   
Uid:    0       0       0       0   
Gid:    0       0       0       0   
FDSize: 256                         
Groups: 0                           
VmPeak:     9436 kB                 
VmSize:     9436 kB                 
VmLck:         0 kB                 
VmHWM:       804 kB                 
VmRSS:       804 kB                 
VmData:      192 kB
VmStk:        84 kB
VmExe:        32 kB
VmLib:      1676 kB
VmPTE:        40 kB
Threads:        1
SigQ:   11/16382
SigPnd: 0000000000000000
ShdPnd: 0000000000004000
SigBlk: fffffffe7ffadeff
SigIgn: 0000000000000000
SigCgt: 0000000180000000
CapInh: 0000000000000000
CapPrm: ffffffffffffffff
CapEff: ffffffffffffffff
CapBnd: ffffffffffffffff
Cpus_allowed:   3
Cpus_allowed_list:      0-1
Mems_allowed:   00000000,00000001
Mems_allowed_list:      0
voluntary_ctxt_switches:        1
nonvoluntary_ctxt_switches:     1

Revision history for this message

Scott James Remnant (Canonical) (canonical-scott) wrote on 2009-09-02:

#22

Thanks, this confirms what we expected.

The TERM signal is being delivered to the process, it's just that the process's mask is set to block that signal and it's still pending.

We're not entirely sure yet what causes this bug; it's certainly not a procps or kernel bug, so I'm reassigning to openssh for the time being since that seems to be the key process.

affects:	procps (Ubuntu) → openssh (Ubuntu)
Changed in openssh (Ubuntu):
status:	Incomplete → Confirmed

Revision history for this message

Ken Bloom (kbloom) wrote on 2009-10-05:

#23

This looks related to http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=549376, as I (the original reporter of that bug) am seeing the same SigBlk mask on my root sshd process. It may also be related to https://bugzilla.mindrot.org/show_bug.cgi?id=271 where sshd was inheriting a bad signal mask from somewhere else.

Revision history for this message

Ken Bloom (kbloom) wrote on 2009-10-05:

#24

System daemons with screwy sigprocmasks on Debian Unstable Edit (1.4 KiB, text/plain)

I'm running Debian Unstable booting with file-rc, and there are several system daemons that have screwy SigBlk masks, of which sshd is one. They are listed in the attached "commands" file.

You can get the data for a similar report on your own machine by running the following commands (as root):

grep SigBlk /proc/*/status | grep -v 0000000000000000 > commands
for x in $(grep SigBlk /proc/*/status | grep -v 000000000000 | sed -e 's@/proc/$.*$/status.*$@\1@g'); do echo -n "$x: "; cat /proc/$x/cmdline; echo; done >> commands

After running this, I cleaned things up a bit in vim to better organize the report.

Perhaps people experiencing this bug on Ubuntu could run a similar report?

Revision history for this message

Ken Bloom (kbloom) wrote on 2009-10-05:

#25

Bingo. The bug is at https://bugs.launchpad.net/ubuntu/+source/udev/+bug/407428 and it's already been fixed in udev.

Ubuntu
openssh package

bad signal mask of ssh sessions

Bug Description

Other bug subscribers

Bug attachments

Remote bug watches

Ubuntuopenssh package

bad signal mask of ssh sessions

Bug Description

Other bug subscribers

Bug attachments

Remote bug watches

Ubuntu
openssh package