[ ssic-linux-Bugs-2782858 ] thread group fails to migrate

"SourceForge.net" <[email protected]> Thu, 30 Apr 2009 03:50:50 +0000
Newsgroups gmane.linux.cluster.ssic.devel
Message-ID <[email protected]>
Bugs item #2782858, was opened at 2009-04-28 09:39
Message generated for change (Comment added) made by rogertsang
You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=405834&aid=2782858&group_id=32541

Please note that this message will contain a full copy of the comment thread,
including the initial issue submission, for this request,
not just the latest update.
Category: Process Management
Group: v1.9.1
Status: Open
Resolution: None
Priority: 5
Private: No
Submitted By: Roger Tsang (rogertsang)
>Assigned to: Roger Tsang (rogertsang)
Summary: thread group fails to migrate

Initial Comment:
[created from discussion http://article.gmane.org/gmane.linux.cluster.ssic.devel/6010]

When running the attached test program the origin node displays the following messages:

add_thread_group:Cannot move process 438181 (calc) - cannot find all shares
add_thread_group: group(438181)/share imbalance 4/1

To build attached test program
# gcc -o calc pthread-test.c -pthread -lm

----------------------------------------------------------------------

>Comment By: Roger Tsang (rogertsang)
Date: 2009-04-29 23:50

Message:
ran into other bugs while fixing this...
1) fails to migrate due to shared SysV sem
2) sending signal after migration results in zombie thread group when the
parent of the test program reside on a different node

----------------------------------------------------------------------

Comment By: Roger Tsang (rogertsang)
Date: 2009-04-28 09:42

Message:
from discussion
http://article.gmane.org/gmane.linux.cluster.ssic.devel/6011

John Huges wrote:
I've also seen another error when trying to migrate the process:

$ migrate 1 395359 
migrate: process 395359 has exited/is exiting
$ dmesg | tail -1
add_thread_group:Cannot move process 394990 (gdm) has different tgid
394990 (expected 395359)

It seems confused about which threads belong to which processes.

I get the same behaviours with a 2.6.11 kernel based on CVS as of 16
March.

----------------------------------------------------------------------

You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=405834&aid=2782858&group_id=32541

------------------------------------------------------------------------------
Register Now & Save for Velocity, the Web Performance & Operations 
Conference from O'Reilly Media. Velocity features a full day of 
expert-led, hands-on workshops and two days of sessions from industry 
leaders in dedicated Performance & Operations tracks. Use code vel09scf 
and Save an extra 15% before 5/3. http://p.sf.net/sfu/velocityconf