Core dump not being generated in RHEL 6.4

khandu · 04-03-2013, 06:52 PM

Hi

New to linux here but after reading a bit..

We have a process being run by "user1". It crashes and we can see a segfault in the /var/log/messages

Other settings of the server are

Quote:

[10:13:28]user1 hostname:~$cat /proc/sys/kernel/core_pattern
/srv/home/user1/core.%p.%e.%t
[10:13:34]user1 hostname:~$cat /proc/sys/kernel/core_uses_pid
1
[10:13:57]user1 hostname:~$ulimit -a
core file size (blocks, -c) unlimited
data seg size (kbytes, -d) unlimited
scheduling priority (-e) 0
file size (blocks, -f) unlimited
pending signals (-i) 127462
max locked memory (kbytes, -l) 64
max memory size (kbytes, -m) unlimited
open files (-n) 1024
pipe size (512 bytes, -p) 8
POSIX message queues (bytes, -q) 819200
real-time priority (-r) 0
stack size (kbytes, -s) 10240
cpu time (seconds, -t) unlimited
max user processes (-u) 1024
virtual memory (kbytes, -v) unlimited
file locks (-x) unlimited

Still we cannot find any core dump file to send to vendor for diagnostics..

How can we fix this??

ShadowCat8 · 04-03-2013, 07:57 PM

Greetings,

Just for a little clarification, when you say that it is crashing, is it locking the system up as well? In other words, are you having to reboot the system to recover from the crash?

Let us know.

khandu · 04-03-2013, 08:43 PM

Quote:

Originally Posted by ShadowCat8

Greetings,

Just for a little clarification, when you say that it is crashing, is it locking the system up as well? In other words, are you having to reboot the system to recover from the crash?

Let us know.

Just the application is freezing / crashing.. System is fine..

Just confused why coredump is not generated with all those settings...

doing a manual coredump via gdb command.. but not core dump automatically does not help when we are not monitoring.

Satyaveer Arya · 04-04-2013, 04:17 AM

khandu,

Checkout these two links, it may help you-

http://honglus.blogspot.in/2010/06/h...-dump-for.html
http://www.cyberciti.biz/tips/linux-core-dumps.html

ShadowCat8 · 04-04-2013, 11:42 AM

Okay,

@Satyaveer Arya: Good info.

And, another thing that you could do to really get to the root cause of the issue would be to use strace.

Running something like this could show you what is happening as the process makes any open, close, read or write system calls:

Code:

 ~ $ strace -o <output_file.txt> -e trace=open,close,read,write -p<process ID#>

This will attach strace to the process you give to the -p option. If you want to capture everything that happens when the program runs, you can use this:

Code:

 ~ $ strace -o <output_file.txt> -e trace=all <program executable> [<any args to the program>]

WARNING: Capturing everything can lead to VERY large trace output files!

And, lastly, to capture *just* the process management type calls (e.g. fork, wait, and exec), you can use the following:

Code:

 ~ $ strace -o <output_file.txt> -e trace=process <program executable> [<any args to the program>]

There are many options and ways to use strace to figure out where trouble may lie in the running of a given program, so definitely check man strace to use the other options if you don't find the core issue with my suggestions above.

HTH. Let us know.

khandu · 04-04-2013, 04:02 PM

Ok.. One more confusion..

I have edited /etc/sysctl.conf to have the following line

Quote:

kernel.core_pattern = /tmp/core.%e.%p

When I execute sysctl -p it changes for the user.. But after reboot for a user if I do

Quote:

$cat /proc/sys/kernel/core_pattern
|/usr/libexec/abrt-hook-ccpp %s %c %p %u %g %t %e

It reverts back.. Although the above line is still in /etc/sysctl.conf..

How do I make this stick for all users.. Don't want to do sysctl -p everytime

Satyaveer Arya · 04-05-2013, 08:28 AM

khandu,

How did you edit the file? Using sysctl -w command or by opening the sysctl.conf file using any editor?