On Fri, Jun 05, 2009 at 02:56:36PM +0800, Tao Ma wrote:No, it is there.
Amerigo Wang wrote:On Fri, Jun 05, 2009 at 02:07:58PM +0800, Tao Ma wrote:in el5, when start kdump service it will do something likeAmerigo Wang wrote:Hmm, my x86_box has 8G mem, the size of kcore looks muchOn Fri, Jun 05, 2009 at 12:03:52PM +0800, Tao Ma wrote:Only 4G.Hi list,Hmm, what is your physical RAM size on test8?
In 2.6.30-rc8, /proc/kcore in x86_64's size is unreasonable large to be 281474974617600.
While in a x86 box, it is 931131392 which looks sane.
[root@test8 ~]# ll /proc/kcore
-r-------- 1 root root 281474974617600 Jun 5 11:15 /proc/kcore
[root@ocfs2-test9 ~]$ ll /proc/kcore
-r-------- 1 root root 931131392 Jun 5 11:58 /proc/kcore
/proc/kcore looks fine on my x86_64 box.
saner than the above huge number, but it is still wrong
according to what the man page describes...
Please do what Andrew said, it will be helpful.
IIRC, kdump should use /proc/vmcore, instead of /proc/kcore...I just started kdump and get the error message.I just noticed this when kexec fails in "Can't find kernel text map area from kcore".It looks like that error message is from userspace?
Is there something wrong?
nothing related.
/sbin/kexec --args-linux -p '--command-line=ro root=LABEL=/ rhgb quiet irqpoll maxcpus=1' --initrd=/boot/initrd-2.6.18-53.el5kdump.img /boot/vmlinuz-2.6.18-53.el5
And the error message is from there.
From /sbin/kexec? I just checked the source code of kexec-tools,I haven't found that message...