IBM BladeCenter Deployment Guide

6.Exit the utility.

7.Press Ctrl+Alt+Backspace.

The next time you log in to the X Window System, the remote mouse will work correctly.

13.Boot-from-SAN fails with Gatekeeper LUN on EMC backed storage – When installing in a boot-from- SAN configuration on an EMC Symmetrix storage array, the following message appears: “The partition table on device sda was unreachable. To create new partitions it must be initialized.” Or after rebooting, you simply see a blinking cursor.

This happens because the Symmetrix uses LUN 0 as a gatekeeper LUN, which is marked as a pseudo-device. The Linux installer maps LUN 0 to /dev/sda. When the installer tries to configure the boot LUN, it is unable to use /dev/sda, so it uses /dev/sdb instead. This leads to a problem with identifying the boot loader.

There are two possible solutions to this problem. One solution is to mask or renumber the LUN of the gatekeeper; the other is to change the boot loader configuration.

1. Masking or Renumbering the Gatekeeper (VCM) LUN

To boot from SAN, ESX Server requires that the boot LUN be the lowest-numbered LUN in the storage array. To accomplish this, you may follow one of two approaches (in order of preference):

To prevent the installer from attempting to write to this device, EMC recommends masking the LUN.

1.Initialize the Volume Logix database using the administrative host.

2.Modify the active configuration file to enable the fba flag2. This restricts access to the Volume Logix database.

The gatekeeper (VCM) LUN can be assigned a number higher than the boot LUN number. Current Fibre Channel adapters can boot only from LUN numbers up to 15, so your boot LUN number must be in the range 0 to 15. Renumber the gatekeeper (VCM) LUN to 16 or higher,

and it won't interfere with the boot LUN.

To renumber the LUN, either contact the vendor for assistance or consult your storage array management documentation.

2.Changing the Boot Loader Configuration

Another way to deal with the gatekeeper LUN is to specify in your boot configuration where to find the boot LUN. This requires editing the /etc/lilo.conf file.

Note: If you have already completed the installation and you are unable to boot the server, you should boot from a Linux rescue CD and mount the boot LUN so you can make the following changes.

1.Back up the file /etc/lilo.conf.

2.Edit the file /etc/lilo.conf.

3.Find the line that says default=esx.

4.After the default line, insert these two lines: disk=/dev/sdb

bios=0x80

Use a tab to indent the second line.

5.Exit the editor, saving your changes.

6.Run lilo.

7.Reboot the server. This time, it should boot successfully from /dev/sdb.

4.6Other Linux Distributions and Blades

We’ve just begun our own testing with SLES 9 and Red Hat 4 on blades in a boot to SAN environment and we hope to add more material here in the future. At this point, the only new problem encountered was that our 8832 blade would not boot from the .iso image of SLES9 mounted via the Management Module. It would boot from the actual CDROM mounted remotely, and the install.

34

http://w3-03.ibm.com/support/techdocs/atsmastr.nsf/WebIndex/WP100564

© 2006 IBM Version 3.0, 3/10/2006

The BladeCenter Deployment Guide is intended as a personal productivity tool. It is not intended to be comprehensive and is provided for

guidance only, on an 'as is basis' without warranty of any kind. Please be aware that its contents have not been certified by IBM.

Page 34
Image 34
McDATA 8832 warranty Other Linux Distributions and Blades, Changing the Boot Loader Configuration