I am trying to install SCP 2.1 BetaDrop1 in my environment. Platform is RHEL 6.0 64bit.
I managed to setup the PXE server and installed 2 storage nodes in my environment.
But the installation of the compute nodes fail. Please see the screenshot for the detailed error message. There seems to be a problem with creating the partition layout?
I have already tried to set "FORCE_PARTITION=yes" in the "/var/lib/tftpboot/pxelinux.cfg/default" file on the firstnode, but the error message stays the same.
I am new to SCP, maybe there are more parameters to change that I am not aware of?
My compute node is a IBM x3850 that is connected to a SAN LUN with about 1 TB free space, no physical disk.
I don't think that the SAN is a problem, because I run into the same error message in a VM that I had booted in the same network.
Do you have any suggestions to fix this?
NOTICE: developerWorks Community will be offline May 29-30, 2015 while we upgrade to the latest version of IBM Connections. For more information, read our upgrade FAQ.
This topic has been locked.
7 replies Latest Post - 2012-07-02T02:39:23Z by Catherine_cherry
Pinned topic Compute Node boot fails
Answered question This question has been answered.
Unanswered question This question has not been answered yet.
Updated on 2012-07-02T02:39:23Z at 2012-07-02T02:39:23Z by Catherine_cherry
rossdavibm 060000884R25 PostsACCEPTED ANSWER
Re: Compute Node boot fails2012-04-13T20:36:38Z in response to DJSchaeferIn checking the system as it boots, does it 'see' a disk of any sort from the BIOS perspective? As you can see in the error, it's looking for a disk /dev/sda, indicating the 'first' disk or a single disk defined on the system. I've seen some situations where external SAS drives (not SAN, I know but a similar concept) only report as 'disk 1' or 'disk 2'; that is, there is no drive 0 (sda) on the system. This means that if the external SAS disk was the only disk, it would be registered to the system as '/dev/sdb' or '/dev/sdc'; not /dev/sda.
Perhaps you have a similar issue? The root question is, does the system see the SAN drive as drive 0?
Re: Compute Node boot fails2012-05-17T17:51:12Z in response to DJSchaeferI have this same issue too. I configured Boot from SAN on my Bladecenter but fails with this error. Initially I installed compute node on local drives (40 GB SSD's) and tried to use SAN partition by configuring "IAASDEV=/dev/mapper/mpathbp1" parameter in "/var/lib/tftpboot/pxelinux.cfg/default" file but it didn't like it.
Wondering what folks would do if you have smaller drives or want to use SAN partition for compute node.
In case of Storage node, I surpassed this issue by commenting out the line
in /data/repos/isaac/storage-ks.cfg on the PXE server.
Re: Compute Node boot fails2012-05-30T15:57:38Z in response to SystemAdminAny updates on this issue?
Is it possible to use linux livecd to boot from multipath device such as SAN?
If not, can we at least make one partition IAASDEV to utilize multipath device (SAN)?
Snow_Cherry 270005DMHM3 PostsACCEPTED ANSWER
Re: Compute Node boot fails2012-06-08T03:12:57Z in response to SystemAdminIt's true it doesn't handle the situation that SAN is the compute node's storage, I wonder if put multipath into the ramdisk, then the first device name will always be /dev/mapper/mpatha. Can you confirm me with this info, then we can change it accordingly.
Catherine_cherry 270005A25C3 Posts
Catherine_cherry 270005A25C3 Posts