[Linux-cluster] Custom Anaconda to see 32 LUNS?

Rick Stevens rstevens at internap.com
Thu Mar 15 18:38:01 UTC 2007


On Thu, 2007-03-15 at 10:41 -0700, Michael Will wrote:
> How does a blade know to boot from a specific lun?
> 
> Can you do a LUN-SAN mapping that would make a lun available to only one
> specific blade each
> under the correct name?

That is an issue.  In those cases, generally you must get onto the
storage device and do host LUN mapping.  Essentially you only expose
the LUN you want to boot from to the WWN of the HBA of the host and hide
it from the other WWNs.

> 
> 
> Michael
> -----Original Message-----
> From: linux-cluster-bounces at redhat.com
> [mailto:linux-cluster-bounces at redhat.com] On Behalf Of
> isplist at logicore.net
> Sent: Thursday, March 15, 2007 11:13 AM
> To: linux-cluster; anaconda-devel-list at redhat.com;
> redhat-install-list at redhat.com
> Subject: [Linux-cluster] Custom Anaconda to see 32 LUNS?
> 
> Is there anyone out there who might have a custom anaconda version which
> could see at least 32 LUNS or more?
> 
> I have yet to find a way of installing blades onto 32 volumes which are
> on a fibre channel storage device. I must find a solution and was now
> told that I might be able to modify anaconda. Not knowing how to do
> this, I wonder if someone might already have a modified version?
> 
> Mike
> 
> This was my old message, things I've done/tried which have not worked.
> 
> //
> I've tried everything I can find and think of or that has been
> suggested.
> 
> I have a Xyratex/MTI type chassis split into 32 volumes. I need to
> install 32 blades onto each volume so that I can remove the drives on
> each blade.
> 
> When I start a Linux install on a blade, it never sees all of the
> volumes, only LUNS 0/1. I need to see all 32 LUNS so that I can install
> all of my servers.
> 
> I've tried all of the following;
> 
> RHEL4, CentOS4.4, others. I don't really care what the distro is, so
> long as it runs basic services such as web/php, qmail, etc. What I do
> need however is that they be GFS/Cluster machines.
> 
> I've tried passing the information at the installers command line;
> 
> scsi_mod.max_luns=256
> scsi_mod.scsi_dev_flags=INLINE:TF200:0x242
> 
> I've tried many variations of these types of commands with no result.
> 
> I've then set up a PXE boot server thinking that I might be able to pass
> the options using pxelinux.cfg. Still no luck, it only sees two LUNS.
> 
> I've tried installing from network with a recompilled initrd.img from a
> machine which was already installed and modprobe.conf modified to see
> the LUNS. The donor server can see all of the volumes, the installing
> version dies with a kernel problem since it cannot see the same already
> installed volume.
> 
> What in the world can I do? Is there a guru here who can tell me how I
> can do this?
> //
> 
> 
> 
> --
> Linux-cluster mailing list
> Linux-cluster at redhat.com
> https://www.redhat.com/mailman/listinfo/linux-cluster
> 
> --
> Linux-cluster mailing list
> Linux-cluster at redhat.com
> https://www.redhat.com/mailman/listinfo/linux-cluster
> 
----------------------------------------------------------------------
- Rick Stevens, Principal Engineer          rstevens at vitalstream.com -
- VitalStream, Inc.                       http://www.vitalstream.com -
-                                                                    -
-         It is better to have loved and lost.  Cheaper, too!        -
----------------------------------------------------------------------




More information about the Linux-cluster mailing list