Description of problem: I tried to setup a network installation with Linux using standard PXE (not the pxe-package that comes with RedHat-8.0) using standard DHCP and TFTP servers. When I called the kernel vmlinuz-8.0 it printed a garbage string on the screen just after the kernel was started (the pxelinux.0 started and parsed the configuration file ok though). Version-Release number of selected component (if applicable): How reproducible: Always Steps to Reproduce: 1. Install standard DHCP and TFTP (I used windows-2000) 2. Place the pxelinux.0 and a kernel named vmlinuz-8.0 3. Perform a PXE-boot of a client. Actual results: Crash Expected results: Normal boot Additional info:
what version of pxelinux? This works fine here - we have plenty of configs with . in them.
Works for me using a Red Hat 8 box as the dhcp+tftp server. Note that some older boot roms require use of the pxe package, whereas others can boot using tftp. Check the pxelinux FAQ for additional information on workarounds needed for various boot rom bugs.
I have used a number of different versions of pxelinux and they all fail the same. The thing is (i forgot to mention that first) is that when booting a RedHat-7.2 kernel (with a dot in the name) it does work. My feeling is that the error is not in syslinux but in the actual kernel.
This is not a bug. A file ending in .0 is assumed to be a PXE NBP and not a Linux kernel file. This is documented in the SYSLINUX/PXELINUX documentation.
I'm closing this, given hpa's comment.