Process remians in Running state causing other similar process to sleep and results to system hang

UNIX for Advanced & Expert Users
1

Hi Experts,

I am facing one problem here which is one process always stuck in running state which causes the other similar process to sleep state . This causes my system in hanged state.

On doing cat /proc/<pid>wchan showing the "__init_begin" in the output.

Can you please help me here why it is getting stuck in the __init_begin API. I am using linux kernel version 2.6.17.

Also i am not facing such probelm in kernel version 2.6.39

i have seen they have given one fix inside __init_begin in the kernel version 2.6.37 . Below is the fix

diff --git a/arch/um/kernel/uml.lds.S b/arch/um/kernel/uml.lds.S
index e7a6cca..664f942 100644
--- a/arch/um/kernel/uml.lds.S
+++ b/arch/um/kernel/uml.lds.S
@@ -22,7 +22,7 @@ SECTIONS
   _text = .;
   _stext = .;
   __init_begin = .;
-  INIT_TEXT_SECTION(PAGE_SIZE)
+  INIT_TEXT_SECTION(0)
   . = ALIGN(PAGE_SIZE);

   .text      :

Now , how this fix related to my problem . please help me to overcome this problem
Looking forward for your suggestions and comments.

2

We need a little more information than that.

What's your system?

What do these processes do?

If this is Linux, 2.6.18 is extremely old. I think that's the first kernel I ever used (and rapidly upgraded from due to bugs).

3

Hi Corona,

I am using linux kernel 2.6.17 .

The process is actually my binary which i am running on switch.

4

Which 'switch' do you mean?

6

Hi

here i am using network packet brokering device.

7

How much memory does your network switch have available when things start failing to run? Devices like this often have very limited resources.

8

Actually -- your kernel is old enough that it's using LinuxThreads instead of modern NPTL. LinuxThreads had an irreparable design flaw which led to its abandonment and replacement by NPTL. Under some circumstances, control signals between processes overflow and are simply lost, causing hanging/stuck processes strikingly similar to this. I had the same problem, way way back in the day. Ancient history now, unless you're using a 15 year old kernel.

You need to upgrade your kernel.