Thursday, August 28, 2014

Trying to allocate 1005 pages for VMLINUZ error when booting with RHEL or CentOS 6.5 disks

I was booting the RHEL 6.5 or CentOS 6.5 on a IBM PureFlex System and I have this error. This occurs when When installing Red Hat Enterprise Linux 6 from DVD media the installation will default to native Extensible Firmware Interface (EFI) mode boot. I do not have

According to IBM Website, 

The workaround is simply to install the operating system in the traditional legacy mode, since there is generally no reason to install in other than Legacy mode. The workaround is only necessary if the media you are booting defaults to EFI mode (DVD or EFI Preboot eXecution Environment (PXE)) otherwise a legacy installation (e.g. - traditional PXE) is the default and is unaffected by this issue.

To force a legacy installation of the operating system from the EFI bootable DVD media the user should:

    Press F12 key when the IBM splash screen is shown during system boot.
    Select Legacy Only option and press Enter.
    The operating system will boot and install in traditional legacy boot mode.

And the issue was resolved.

References:
  1. Red Hat Enterprise Linux 6 (RHEL6) native Extensible Firmware Interface (EFI) install is not supported with greater than 512 GB memory - IBM System x and BladeCenter
  2. Bug 691860 - UEFI version of ISO fails to boot when >4gig (since f14)

Wednesday, August 27, 2014

Location of GPFS Client Log file

The location of the GPFS Log file is located at the /var/adm/ras/mmfs.log.latest. You can find a wealth of information of errors and information. When you are monitoring the errors in real time, you may want to use the tail -f to see the real time troubleshooting

# tail -f /var/adm/ras/mmfs.log.latest

Monday, August 25, 2014

Accurately Time Your Parallel Loops in OpenMP

Learn how to use a wall clock timer that is included as part of the OpenMP library. Jeff Cogswell walks you through how. Taken from Accurately Time Your Parallel Loops in OpenMP which is found http://goparallel.sourceforge.net/

Sunday, August 24, 2014

GPFS Autoconfig Error on CentOS 6

When I try to build the portability layer for GPFS which is done by

make LINUX_DISTRIBUTION=REDHAT_AS_LINUX Autoconfig
 There are errors like the ones below
cd /usr/lpp/mmfs/src/config; ./configure --genenvonly; /usr/bin/cpp -P def.mk.proto > ./def.mk; exit $? || exit 1
Cannot find a valid kernel include dir
def.mk.proto:12:19: error: env.mcr: No such file or directory
make: *** Autoconfig Error 1

The solution is to make sure you have all the package installed. These are the following prerequistics

kernel-headers-2.6.18-128.el5.x86_64.rpm
kernel-devel-2.6.18-128.el5.x86_64.rpm
libgomp-4.3.2-7.el5.x86_64.rpm
glibc-headers-2.5-34.x86_64.rpm
glibc-devel-2.5-34.x86_64.rpm
gcc-4.1.2-44.el5.x86_64.rpm
libstdc++-devel-4.1.2-44.el5.x86_64.rpm
gcc-c++-4.1.2-44.el5.x86_64.rpm

References:
  1. Linux system does not have prerequisite kernel or C++ packages 

Saturday, August 23, 2014

Accurately Time Your Parallel Loops in OpenMP

There is an interesting video by Go Parallel Site on "Accurately Time Your Parallel Loops in OpenMP".
I have embedded the URL here