Home » General » Support » ovzkernel-xen 4gb fix up issue
| ovzkernel-xen 4gb fix up issue [message #34537] |
Fri, 16 January 2009 04:34  |
victorskl
Messages: 28
Registered: September 2006
|
Junior Member |
|
|
Hi,
I found when running ovzkernel-xen based kernel, 4gb fix up issue could not solve out. Here are a few report.
[root@firefly ~]# yum list installed | grep kernel
kernel-headers.i386 2.6.18-92.1.22.el5 installed
kernel-xen.i686 2.6.18-92.1.22.el5 installed
kernel-xen.i686 2.6.18-92.1.13.el5 installed
kernel-xen.i686 2.6.18-92.el5 installed
kernel-xen-devel.i686 2.6.18-92.1.22.el5 installed
ovzkernel-xen.i686 2.6.18-92.1.13.el5.028 installed
[root@firefly ~]# yum list installed | grep glibc
glibc.i686 2.5-24.el5_2.2 installed
glibc-common.i386 2.5-24.el5_2.2 installed
glibc-devel.i386 2.5-24.el5_2.2 installed
glibc-headers.i386 2.5-24.el5_2.2 installed
Note that, each kernel-xen created `hwcap 0 nosegneg` bit, since all are rhel base kernels and red hat include this xen bit patch.
[root@firefly ~]# ls /etc/ld.so.conf.d/
kernelcap-2.6.18-92.1.13.el5.028stab059.6.conf kernelcap-2.6.18-92.el5.conf
kernelcap-2.6.18-92.1.13.el5.conf mysql-i386.conf
kernelcap-2.6.18-92.1.22.el5.conf
[root@firefly ~]# cat /etc/ld.so.conf.d/kernelcap-*
# This directive teaches ldconfig to search in nosegneg subdirectories
# and cache the DSOs there with extra bit 0 set in their hwcap match
# fields. In Xen guest kernels, the vDSO tells the dynamic linker to
# search in nosegneg subdirectories and to match this extra hwcap bit
# in the ld.so.cache file.
hwcap 0 nosegneg
# This directive teaches ldconfig to search in nosegneg subdirectories
# and cache the DSOs there with extra bit 0 set in their hwcap match
# fields. In Xen guest kernels, the vDSO tells the dynamic linker to
# search in nosegneg subdirectories and to match this extra hwcap bit
# in the ld.so.cache file.
hwcap 0 nosegneg
# This directive teaches ldconfig to search in nosegneg subdirectories
# and cache the DSOs there with extra bit 0 set in their hwcap match
# fields. In Xen guest kernels, the vDSO tells the dynamic linker to
# search in nosegneg subdirectories and to match this extra hwcap bit
# in the ld.so.cache file.
hwcap 0 nosegneg
# This directive teaches ldconfig to search in nosegneg subdirectories
# and cache the DSOs there with extra bit 0 set in their hwcap match
# fields. In Xen guest kernels, the vDSO tells the dynamic linker to
# search in nosegneg subdirectories and to match this extra hwcap bit
# in the ld.so.cache file.
hwcap 0 nosegneg
When running with RHEL5.2 latest 22 kernel, it run fine. No error printk. libc.so.6 point to nosegneg.
[root@firefly ~]$ uname -r
2.6.18-92.1.22.el5xen
[root@firefly ~]$ ldd /sbin/init
linux-gate.so.1 => (0x00a01000)
libsepol.so.1 => /lib/libsepol.so.1 (0x00b99000)
libselinux.so.1 => /lib/libselinux.so.1 (0x00b7f000)
libc.so.6 => /lib/i686/nosegneg/libc.so.6 (0x00110000)
libdl.so.2 => /lib/libdl.so.2 (0x00b37000)
/lib/ld-linux.so.2 (0x009d1000)
When running with RHEL5 13 kernel(where ovzkernel-xen 2.6.18-92.1.13.el5.028stab059.6xen base off), it also run fine. No error printk. libc.so.6 point to nosegneg.
[root@firefly ~]# uname -r
2.6.18-92.1.13.el5xen
[root@firefly ~]# ldd /sbin/init
linux-gate.so.1 => (0x00265000)
libsepol.so.1 => /lib/libsepol.so.1 (0x00b99000)
libselinux.so.1 => /lib/libselinux.so.1 (0x00b7f000)
libc.so.6 => /lib/i686/nosegneg/libc.so.6 (0x009ef000)
libdl.so.2 => /lib/libdl.so.2 (0x00b37000)
/lib/ld-linux.so.2 (0x009d1000)
Once I installed ovzkernel-xen and boot on, it fill the console with 4gb fix up printk messages. libc.so.6 _do not_ point to nosegneg, anymore.
[root@firefly ~]# uname -r
2.6.18-92.1.13.el5.028stab059.6xen
[root@firefly ~]# ldd /sbin/init
libsepol.so.1 => /lib/libsepol.so.1 (0x00b99000)
libselinux.so.1 => /lib/libselinux.so.1 (0x00b7f000)
libc.so.6 => /lib/libc.so.6 (0x00a1e000)
libdl.so.2 => /lib/libdl.so.2 (0x00110000)
/lib/ld-linux.so.2 (0x009d1000)
As soon as node boot up and start init pid 1, then it started 4gb fix up printk.
| Quote:Console Messages |
......
......
Setting up other filesystems.
Setting up new root fs
no fstab.sys, mounting internal defaults
Switching to new root and running init.
unmounting old /dev
unmounting old /proc
unmounting old /sys
4gb seg fixup, process init (pid 1), cs:ip 73:004ae5a0
4gb seg fixup, process init (pid 1), cs:ip 73:00524939
4gb seg fixup, process init (pid 1), cs:ip 73:004b16cc
4gb seg fixup, process init (pid 1), cs:ip 73:004b16da
4gb seg fixup, process init (pid 1), cs:ip 73:004b16cc
4gb seg fixup, process init (pid 1), cs:ip 73:004b16da
4gb seg fixup, process init (pid 1), cs:ip 73:00469a8b
4gb seg fixup, process init (pid 1), cs:ip 73:00469a74
4gb seg fixup, process init (pid 1), cs:ip 73:004b77e5
4gb seg fixup, process init (pid 1), cs:ip 73:00469a74
INIT: version 2.86 booting
Welcome to CentOS release 5.2 (Final)
Press 'I' to enter interactive startup.
printk: 115503 messages suppressed.
4gb seg fixup, process hwclock (pid 766), cs:ip 73:0014b810
......
......
<snip>
......
4gb seg fixup, process sendmail (pid 4343), cs:ip 73:00ce06cc
printk: 16 messages suppressed.
4gb seg fixup, process sendmail (pid 4343), cs:ip 73:00ce06cc
printk: 18 messages suppressed.
4gb seg fixup, process sendmail (pid 4343), cs:ip 73:00ce06cc
printk: 16 messages suppressed.
4gb seg fixup, process sendmail (pid 4343), cs:ip 73:00ce06cc
printk: 16 messages suppressed.
4gb seg fixup, process sendmail (pid 4343), cs:ip 73:00ce06cc
printk: 15 messages suppressed.
|
The error is reproducible. I also found others' happening on this list thread.
http://openvz.org/pipermail/users/2008-December/002462.html
I've tried to rebuild from http://wiki.openvz.org/Download/kernel/rhel5/028stab059.6 source rpm but still not work. I found the following threads might give some help. It would be related to OpenVZ code to align with how Glibc treat on Xen's TLS/ELF segment, regarding syscall to user space linker _but_ not so sure if my understanding is correct. I've tried to check in ovzkernel code but still new in kernel hacking and openvz code is a big patch also, diff about 800++ files on rhel base.
http://lists.xensource.com/archives/html/xen-devel/2005-11/m sg00262.html
http://lists.xensource.com/archives/html/xen-devel/2005-11/m sg00078.html
http://lkml.indiana.edu/hypermail/linux/kernel/0704.3/0002.h tml
https://lists.linux-foundation.org/pipermail/virtualization/ 2007-May/007631.html
Please kindly help prior on this issue, as message logs is growing up and kind of effecting in performance, if keep stubborn to run with this kernel.
[root@node ~]# ls -alh /var/log/messages*
-rw------- 1 root root 15M Jan 16 12:39 /var/log/messages
-rw------- 1 root root 19M Jan 11 04:02 /var/log/messages.1
-rw------- 1 root root 9.2M Jan 4 04:02 /var/log/messages.2
-rw------- 1 root root 34K Dec 28 04:02 /var/log/messages.3
-rw------- 1 root root 294K Dec 21 04:02 /var/log/messages.4
Thanks.
P.S.. This regardless of Dom0(Host) or DomU(Guest), which ever running this kernel will get printk!
[Updated on: Fri, 16 January 2009 10:19] Report message to a moderator |
|
|
|
| Re: ovzkernel-xen 4gb fix up issue [message #44854 is a reply to message #34537] |
Mon, 09 January 2012 07:20  |
apis17
Messages: 1
Registered: January 2012
|
Junior Member |
|
|
it fixed using i386 version of glibc
[root@linux ~]#yumdownloader glibc.i386
then yum downloading 'glibc-2.5-65.i386.rpm' package
[root@linux ~]#rpm -Uvh --force glibc-2.5-65.i386.rpm
Then check i386 installation
[root@linux ~]# rpm -q glibc.i386
glibc-2.5-65
[root@linux ~]# rpm -q glibc.i686
package glibc.i686 is not installed
maybe this thread is old but this solved 4gb problem on mine.
|
|
|
|
Goto Forum:
Current Time: Thu Oct 17 19:31:32 GMT 2024
Total time taken to generate the page: 0.05350 seconds
|
OpenVZ Forum
Members
Search
Help
Register
Login
Home
