[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Wg-test-framework] Xen Project Question - Chainloading



Steve,
        Let me know if this is what is needed.

Re: UEFI boot
        I have used a separate tftp server than my DHCP server (next-server)
        This commonly is the Ubuntu net installer.  This will go through set up 
options and then go to Ubuntu over the 10GbE to get the installer and packages.

        The installers commonly will partition the SSD/HDD to include the 
needed EFI partition and install the needed 'efi' program (often called 
grubarch64.efi).  They will 'activate' the partition for EFI boot via the AMI 
BIOS so the next boot will be that partition.

        (in other words... about what is expected).

        Also, I have:
                - Used vMedia to virtually mount an ISO and booted/installed
                - Used IPMI SOL to interface the console
                - Used the BMC console to access console

        My peers have done installs using USB.  I have the infrastructure so 
that I have not tried that yet.

-------------------------------------------------------------------------------------------------------------
Frank Motta
Field Applications Engineer for Cavium
408-943-7615 (O)  510-757-8909 (C) 
Phoenics Electronics

www.phoenicselectronics.com 
Proudly celebrating over 10 years supporting Cavium customers

-----Original Message-----
From: Steve Wilson 
Sent: Monday, February 13, 2017 1:48 PM
To: Frank Motta <frank@xxxxxxxxxxxxxxxxxxxxxxx>
Cc: Ian Jackson <ian.jackson@xxxxxxxxxxxxx>; Lars Kurth 
<lars.kurth@xxxxxxxxxx>; Alena Hyde <alena@xxxxxxxxxxxxxxxxxxxxxxx>; 
ypatel@xxxxxxxxxxx; Julien Grall <julien.grall@xxxxxxx>; 
wg-test-framework@xxxxxxxxxxxxxxxxxxxx
Subject: Xen Project Question - Chainloading 

Hi Frank,    As I mentioned on the phone,  Ian provided clarity on his 
chainloading question.....


> Sorry about the opaque terminology.

> What I mean is that (with appropriate configuration) exiting from a 
> UEFI bootloader (acquired via tftp according to a dhcp next-server
> field) causes the firmware to try the next boot option (which in our 
> case will be to boot from the hard disk).
> 
> This allows us to control whether the machine boots the OS on the hard 
> disk by adjusting the configuration files for the network bootloader 
> (grub), without having to execute code on the host.  (This is 
> necessary because a previous test might have loaded code onto the host 
> hard disk which prevents a hard disk boot.)
> 
> We have seen one hardware/firmware combination where exiting the UEFI 
> grub bootloader causes the host to crash.  Would be regarded by you as 
> a firmware bug ?

Pls comment.  Does this work?

Thanks, Steve




-----Original Message-----
From: Frank Motta
Sent: Friday, February 10, 2017 12:31 PM
To: Ian Jackson <ian.jackson@xxxxxxxxxxxxx>; Steve Wilson 
<steve@xxxxxxxxxxxxxxxxxxxxxxx>; Lars Kurth <lars.kurth@xxxxxxxxxx>; Alena Hyde 
<alena@xxxxxxxxxxxxxxxxxxxxxxx>; ypatel@xxxxxxxxxxx; Julien Grall 
<julien.grall@xxxxxxx>; wg-test-framework@xxxxxxxxxxxxxxxxxxxx
Subject: RE: [Wg-test-framework] ThunderX machines (for the Xen Project OSSTEST 
Test Lab) [and 1 more messages]

Ian,
        In reading through this, it appears that you are having OS boot issues.
        Can you provide me with:
        - the complete output from the serial console (the DB9 connector above 
the VGA) - from cold power on until the failure
        - a screen capture of the BMC WebUI which contains the version 
information


        Which version(s) OS are you booting natively and which are the guests?



-------------------------------------------------------------------------------------------------------------
Frank Motta
Field Applications Engineer for Cavium
408-943-7615 (O)  510-757-8909 (C)
Phoenics Electronics

www.phoenicselectronics.com
Proudly celebrating over 10 years supporting Cavium customers

-----Original Message-----
From: Ian Jackson [mailto:ian.jackson@xxxxxxxxxxxxx]
Sent: Thursday, February 9, 2017 5:23 AM
To: Steve Wilson <steve@xxxxxxxxxxxxxxxxxxxxxxx>; Lars Kurth 
<lars.kurth@xxxxxxxxxx>; Alena Hyde <alena@xxxxxxxxxxxxxxxxxxxxxxx>; 
ypatel@xxxxxxxxxxx; Frank Motta <frank@xxxxxxxxxxxxxxxxxxxxxxx>; Julien Grall 
<julien.grall@xxxxxxx>; wg-test-framework@xxxxxxxxxxxxxxxxxxxx
Subject: Re: [Wg-test-framework] ThunderX machines (for the Xen Project OSSTEST 
Test Lab) [and 1 more messages]

Ian Jackson writes ("Re: [Wg-test-framework] ThunderX machines (for the Xen 
Project OSSTEST Test Lab)"):
> Right.  OK.  That is fine then.

I think in terms of my technical questions, then, that leaves only
this:

Ian Jackson writes ("Re: [Wg-test-framework] ThunderX machines (for the Xen 
Project OSSTEST Test Lab)"):
> Sorry about the opaque terminology.
> 
> What I mean is that (with appropriate configuration) exiting from a 
> UEFI bootloader (acquired via tftp according to a dhcp next-server
> field) causes the firmware to try the next boot option (which in our 
> case will be to boot from the hard disk).
> 
> This allows us to control whether the machine boots the OS on the hard 
> disk by adjusting the configuration files for the network bootloader 
> (grub), without having to execute code on the host.  (This is 
> necessary because a previous test might have loaded code onto the host 
> hard disk which prevents a hard disk boot.)
> 
> We have seen one hardware/firmware combination where exiting the UEFI 
> grub bootloader causes the host to crash.  Would be regarded by you as 
> a firmware bug ?

Thanks,
Ian.

_______________________________________________
Wg-test-framework mailing list
Wg-test-framework@xxxxxxxxxxxxxxxxxxxx
https://lists.xenproject.org/cgi-bin/mailman/listinfo/wg-test-framework

 


Rackspace

Lists.xenproject.org is hosted with RackSpace, monitoring our
servers 24x7x365 and backed by RackSpace's Fanatical Support®.