Skip to contentSkip to conversation

proc_stack for x86_64 (spinning process)

DISCUSSION Nuno Pais Fernandes 3 participants 4 messages Apr 8, 2008

lore
Share & tools
lore
#1 lore raw
Quoted 5 lines · Nuno Pais Fernandes

Hi,

I have a spinning process that it's locking my server (eating 100% CPU). I can't kill it even with kill -9

Quoted 7 lines

I'm trying to use proc_stack to check what's the problem, compiling the kernel with

http://linuxhacker.ru/~nikita/patches/2.6.12-rc6/2005.06.11/vm_07-proc-stack.patch

But this patch is only for x86 as my arch is x86_64. Does anyone know if there's a patch for x86_64?

The standard way to handle that is to just do

echo 1 > /proc/sys/kernel/sysrq
echo t > /proc/sysrq-trigger

and then look for the process backtrace in the kernel log.

Another alternative is to use crash to take a live dump and then analyse that using crash.

-Andi

#2 lore raw
Quoted 15 lines · On Tuesday 08 April 2008 12:21:53 Andi Kleen wrote:
Quoted 5 lines · Nuno Pais Fernandes

Hi,

I have a spinning process that it's locking my server (eating 100% CPU). I can't kill it even with kill -9

Quoted 8 lines

I'm trying to use proc_stack to check what's the problem, compiling the kernel with

http://linuxhacker.ru/~nikita/patches/2.6.12-rc6/2005.06.11/vm_07-proc-st
ack.patch

But this patch is only for x86 as my arch is x86_64. Does anyone know if there's a patch for x86_64?

Quoted 6 lines

The standard way to handle that is to just do

echo 1 > /proc/sys/kernel/sysrq
echo t > /proc/sysrq-trigger

and then look for the process backtrace in the kernel log.

I've already done that but the problem is that the "R" script does not show any backtrack. Apparently it's spinning inside kernel and all other processes become in D state (uninterrutable sleep) because of that one.

Apr  7 15:22:03 app14 kernel: comentarios   R  running task       0  7615   
8051                     (NOTLB)
Apr  7 15:22:03 app14 kernel: comentarios   D ffffff80e2570030     0  8150   
7328                     (NOTLB)
Apr  7 15:22:03 app14 kernel: ffffff80cc8bfcf8 0000000000000282 
00000000f69539c0 ffffff80eea385b0
Apr  7 15:22:03 app14 kernel:        ffffff80e2570030 000000000011d7be 
001769ecf4f60e59 ffffff8000e26030
Apr  7 15:22:03 app14 kernel:        ffffff80e25702c8 0000000000000000
Apr  7 15:22:03 app14 kernel: Call 
Trace:<ffffffffa011d4fd>{:ocfs2:ocfs2_dentry_revalidate+369}
Apr  7 15:22:03 app14 kernel:        <ffffffff8018149e>{do_lookup+419} 
<ffffffff80292e6b>{__down+147}
Apr  7 15:22:03 app14 kernel:        
<ffffffff8012b177>{default_wake_function+0} 
<ffffffff80187f0a>{lease_alloc+96}
Apr  7 15:22:03 app14 kernel:        <ffffffff802945cf>{__down_failed+53} 
<ffffffff801c9faf>{dummy_inode_permission+0}
Apr  7 15:22:03 app14 kernel:        
<ffffffff801c9faf>{dummy_inode_permission+0} 
<ffffffff80174017>{.text.lock.open+5}
Apr  7 15:22:03 app14 kernel:        <ffffffff8018ba2e>{__d_lookup+287} 
<ffffffff80181096>{get_write_access+62}
Apr  7 15:22:03 app14 kernel:        <ffffffff80182fd4>{may_open+481} 
<ffffffff80183340>{open_namei+788}
Apr  7 15:22:03 app14 kernel:        <ffffffff80173c22>{filp_open+80} 
<ffffffff8018a91d>{dput+56}
Apr  7 15:22:03 app14 kernel:        <ffffffff801e8195>{strncpy_from_user+74} 
<ffffffff8011f62c>{sys32_open+54}
Apr  7 15:22:03 app14 kernel:        <ffffffff8011d178>{ia32_sysret+0}

Thanks,
Nuno Fernandes

Quoted 4 lines

Another alternative is to use crash to take a live dump and then analyse that using crash.

-Andi

#3 lore raw
Quoted 18 lines · Nuno Fernandes
Quoted 17 lines · On Tuesday 08 April 2008 12:21:53 Andi Kleen wrote:
Quoted 5 lines · Nuno Pais Fernandes

Hi,

I have a spinning process that it's locking my server (eating 100% CPU). I can't kill it even with kill -9

Quoted 8 lines

I'm trying to use proc_stack to check what's the problem, compiling the kernel with

http://linuxhacker.ru/~nikita/patches/2.6.12-rc6/2005.06.11/vm_07-proc-st
ack.patch

But this patch is only for x86 as my arch is x86_64. Does anyone know if there's a patch for x86_64?

The standard way to handle that is to just do

Quoted 7 lines
Quoted 4 lines

echo 1 > /proc/sys/kernel/sysrq
echo t > /proc/sysrq-trigger

and then look for the process backtrace in the kernel log.

I've already done that but the problem is that the "R" script does not show any backtrack. Apparently it's spinning inside kernel and all other processes become in D state (uninterrutable sleep) because of that one.

Then you can either try to catch it with sysrq-p (but only works rarely) or better use the crash method. But if the process is running live dump might not work, you might need to force a non live dump using sysrq-c

Another way to identify spinning code is to use a profiler like oprofile or readprofile. Something that spins tends to be on top.

-Andi