Bug 1421656

Hello, when you change boot order in VMWare to boot from first drive instead network, does it boot?

In "RHV4" If we manually, change the boot order for the provisioned host
from 

1) Network Pxe , 2) Harddisk  to 
1) Harddisk 2)None

The host successfully boot's up from local disk.

There is a bug in libvirt that prevents from booting via PXELinux localboot option. Can you change back to boot from network but before booting the VM edit

/var/lib/tftpboot/pxelinux.cfg/01-MAC_ADDRESS

and change the content to (taken from https://github.com/theforeman/community-templates/blob/e6f8f14bd93e7fb57ce4718b869aec1326898f13/provisioning_templates/snippet/pxelinux_chainload.erb)

LABEL local
  MENU LABEL Chainload into bootloader on the first disk - alternative
  COM32 chain.c32
  APPEND hd0

Make sure the chain.c32 file is present in /var/lib/tftpboot/ if not copy it from /usr/share/syslinux.

Try again. Does it work?

Which RHEL version is this?

Yes, after changing the /var/lib/tftpboot/pxelinux.cfg/01-MAC_ADDRESS file to 

DEFAULT menu
PROMPT 0
MENU TITLE PXE Menu
TIMEOUT 200
TOTALTIMEOUT 6000
ONTIMEOUT local

LABEL local
	MENU LABEL Chainload into bootloader on the first disk - alternative
	COM32 chain.c32
	APPEND hd0

It work's, the host boot's up using the local disk.

This is RHEL7.3 host provisioned via satellite on rhv4

Ok thanks, so the alternative fix for problematic libvirt driver in RHEV 4.0 is to backport this change:

https://github.com/theforeman/community-templates/commit/9769dc89ec5d93193b059aeec3f3f9c80bc160d3

This adds alternative loading menu option to load from disk for these kind of hosts.

This is NOT a regression, RHEV 3.x should work fine I assume. Bryan, the change is trivial (just a template change), but if it is too late, I think we can pull this out to the next errata.

Patch ready, very trivial template update that adds new (disabled by default) entry. Zero risk.

https://access.redhat.com/solutions/2933681

Adding release note request.

I have Sat6.2.6 and when trying to localboot 0 for BL2x220 G5 blades I get illegal opcode.  I changed the template "PXELinux default local boot" I still had to manually edit the /tftpboot/pxelinux.cfg/01-00-17-a4-77-00-ac file to change it to:

label local
MENU LABEL (local)
MENU DEFAULT
COM32 chain.c32
APPEND hd0 0

And then the server booted correctly to the local disk from PXE.

Is there a way to add an alternative entry for local boot into the template?  Here is the foreman bug:

http://projects.theforeman.org/issues/6725

Hello Steven!

A change has been pushed into community-templates repository [1] which adds an entry called "Chainload the first hard drive (hd0)" which uses chain.c32 method. You can make it a default easily via "MENU DEFAULT" and "DEFAULT" options for all of your systems, unfortunately there is no way of changing that per host as I believe @host variable is not available during template rendering (I haven't verified you can try).

This change will be available in Foreman 1.15 and Satellite 6.3. You can easily do the template change in older versions tho. If you want to avoid manual config file edits on your TFTP server, you can click on Cancel build and then Build which will trigger templates redeployment

[1] https://github.com/theforeman/community-templates/commit/23accd5205333a7187d8d78aab5dfb49dd0eea62

hey guys, aren't you just hitting the following bug?

https://bugzilla.redhat.com/show_bug.cgi?id=1467925

if so, ..well it's a known issue not related to rhv. and you can workaround it e.g. by waiving the build by clicking 'cancel build' - this will actually keep the host AR and change its status to BUILT + flips the bootloader to make the host automatically boot from local disk.

This issue is something different and related to chainloading the bootloader for RHV 4.

More details:
https://bugzilla.redhat.com/show_bug.cgi?id=1421656#c4 for more clarification

Build : Satellite 6.3 Snap 18

Failing this as the issue is reproducible

(In reply to Roman Plevka from comment #16)
> hey guys, aren't you just hitting the following bug?
> 
> https://bugzilla.redhat.com/show_bug.cgi?id=1467925
> 
> if so, ..well it's a known issue not related to rhv. and you can workaround
> it e.g. by waiving the build by clicking 'cancel build' - this will actually
> keep the host AR and change its status to BUILT + flips the bootloader to
> make the host automatically boot from local disk.

Roman,

My recommended boot method when using Sat6 is always make net first boot and let Sat6 figure it out so I never need to touch the BIOS.

(In reply to Sanket Jagtap from comment #17)
> This issue is something different and related to chainloading the bootloader
> for RHV 4.
> 
> More details:
> https://bugzilla.redhat.com/show_bug.cgi?id=1421656#c4 for more clarification

Is there some reason why the chain loading should not be the default in all Sat6 templates as this seems to be an issues for physical and VM systems?

(In reply to Sanket Jagtap from comment #17)
> This issue is something different and related to chainloading the bootloader
> for RHV 4.
> 
> More details:
> https://bugzilla.redhat.com/show_bug.cgi?id=1421656#c4 for more clarification

Is there some reason why the chain loading should not be the default in all Sat6 templates as this seems to be an issues for physical and VM systems?

Steven, good question, I just made the change upstream. This will be possible to change via Global setting as well: https://github.com/theforeman/foreman/pull/4891

Rich, back to you. Can we push to GA blocker?

I've updated https://access.redhat.com/solutions/2933681 to note about the need for cloning the template when editing. It should be ready to be used for 6.3

Just for the record, we have patches upstream to make this a seamless experience. These are the patches:

https://github.com/theforeman/foreman/pull/4891
https://github.com/theforeman/community-templates/pull/430

This adds new Global settings "Default PXE global template entry" and "Default PXE local template entry" so users can easily change this WITHOUT playing around with templates. It's pending review. These are small changes and I think they could be backported into either GA or z-stream.

Thank you for your interest in Satellite 6. We have evaluated this request, and we do not expect this to be implemented in the product in the foreseeable future. We are therefore closing this out as WONTFIX. If you have any concerns about this, please feel free to contact Rich Jerrido or Bryan Kearney. Thank you.

The needinfo request[s] on this closed bug have been removed as they have been unresolved for 500 days