This bug is against xCAT 2.9.1 with RHEL 7. While use xCAT to deploy a RHEL 7 diskfull compute node, if there is no hard disk drive attached directly to the compute node, the end user will get the following error messages from the operation system installer, anaconda.
08:12:54 Running pre-installation scripts
The following problem occured on line 1 of the kick start file:
Option --ondisk: invalid string value: '--fstype'
The error message above shows a synatex error. It cost me around 15 minute to realize that there is no local hard disk drive attached. xCAT should give end user a more friendly error message.
IMHO, there are two options.
xCAT try to detect the hard disk configuration and provide a error message to the end user directly.
Just fill /dev/sda to the kickstart file if no disk was found. And anaconda will give an error message to the end user.
Enclosed please find a screen shot for the anaconda error messages mentioned above.
The fix for RHEL could be: if could not determine the install disk, try the default /dev/sda
Retrieving http://10.2.1.2:80/install/rhels6.6/x86_64/images/product.img...
Retrieving http://10.2.1.2:80/install/rhels6.6/x86_64/images/install.img...
Running anaconda 13.21.229, the Red Hat Enterprise Linux system installer - please wait.
Examining storage devices
In progress
Command line mode requires all choices to be specified in a kickstart configuration file.
No disks found
No usable disks have been found.
('custom', ['_Try again', '_Exit installer'])
Last edit: Guang Cheng Li 2015-04-15
On SLES: the error is " Error while configuring partitions.", should be clear enough.
On Ubuntu, the error message is " No root file system, Please correct this from the partitioning menu.", also acceptable...
Fixed in xCAT 2.10
commit bbcfef36e2469f6aa8b0e34698a0649313f42afe
Will close this one. It was fixed for a long time.