Latest LQ Deal: Linux Power User Bundle
Go Back > Forums > Linux Forums > Linux - General
User Name
Linux - General This Linux forum is for general Linux questions and discussion.
If it is Linux Related and doesn't seem to fit in any other forum then this is the place.


  Search this Thread
Old 05-09-2005, 10:21 AM   #1
LQ Newbie
Registered: May 2005
Posts: 3

Rep: Reputation: 0
Core file creation problem on ramdisk system

My problem is a hanging (non responsive) system when a core file is about
to be created on a ramdisk booted X86 SLES SP1 platform.

The ramdisk is small (ramdisk_size=70000) and the size of the core
file is a lot bigger than that. It does not matter if the application
produces the core due to e.g. a SIGSEGV or if it is forced by e.g.
assert() or abort() calls. On top of this there is a /export tmpfs filesystem
which is the "working" directory of the application and the directory
where the core file is actually placed. Everything works fine if
"ulimit -c" is set to 0 or less size than the ramdisk. What is interesting is
that creating a file in /export using dd or similar that is equal the
real size of the core seems to work without problem. So, then why
does the system hang when creating the core file?
Is the core file temporarily created on / or any other place that is
affected by the ramdisk size before moved to target folder?
It at least seems to me that the ramdisk size is the limiting factor
for the kernel core dump file creation routine!?
A similiar (read not identical kernel modules) system booted from HDD works fine I might add

Is there some way of overcoming this issue ( the ramdisk can _not_
be increased in size ) or is this a bug that have no current known
workarounds. Is this in fact a bug or is there a vital kernel module
missing? Any hints or ideas what might be wrong?
Old 05-12-2005, 03:06 PM   #2
LQ Newbie
Registered: May 2005
Posts: 3

Original Poster
Rep: Reputation: 0
Well, problems solved. It was not unique to the ramdisk configuration after all,
the problems where there also at HDD boot.
The problem was isolated to a device driver that did mmap() with only VM_RESERVED
which fooled the core dump handler from SLES SP1 (2.6.5) to try to dump H/W
protected areas. By also doing mmap() using VM_IO solved it. But, there seems to
be issues regarding how the core dump handler is implementing this. If you
compare e.g. SLES (2.6.5) and debian (2.6.11 I think) they have very different
implementations/views of this. In debian all of VM_IO, VM_RESERVED and VM_SHARED
are refused to be dumped to core but in SLES it is only VM_IO that is fully stopped.
If this is relatated to distros or if it is actually the kernel version that are so
completely different I have not yet investigated. Next on schedule I guess


Thread Tools Search this Thread
Search this Thread:

Advanced Search

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off

Similar Threads
Thread Thread Starter Forum Replies Last Post
Building a file system ( Ramdisk) ayoub Linux - Newbie 1 11-10-2005 08:48 PM
Files removed from file system during Fedora Core Upgrade gryphonavocatio Linux - General 5 09-21-2005 12:27 PM
Problem with the creation of new users on my MEPIS system. duffmckagan MEPIS 2 06-11-2005 08:46 PM
strange file creation problem dizzutch Programming 1 10-06-2004 05:30 PM
php read file creation, modified date problem antony_csf Programming 3 08-18-2004 06:46 AM > Forums > Linux Forums > Linux - General

All times are GMT -5. The time now is 01:59 PM.

Main Menu
Write for LQ is looking for people interested in writing Editorials, Articles, Reviews, and more. If you'd like to contribute content, let us know.
Main Menu
RSS1  Latest Threads
RSS1  LQ News
Twitter: @linuxquestions
Facebook: linuxquestions Google+: linuxquestions
Open Source Consulting | Domain Registration