FYI. This is with b126 forward.... MRJ http://www.opensolaris.org/os/community/on/flag-days/pages/2009101201/ = Date: Mon, 12 Oct 2009 13:28:23 +0100 From: Gary Pennington To: onnv-gate Subject: Heads up: Introduction of new xVM SMF milestone You can safely ignore this message if you don''t use the xVM hypervisor on x86 systems. The putback of: 6885672 Provide hypervisor configuration panel and associated service introduces a new SMF milestone, svc:/milestone/xvm, which can be used to conveniently enable/disable the xVM hypervisor. There is no need to use the existing xVM services if this milestone is used. To enable the hypervisor: svcadm enable svc:/milestone/xvm This will use bootadm(1M) to convert an existing menu.lst entry so that on next reboot the hypervisor will be started. The store service, svc:/system/xvm/store, goes into maintenance to indicate that a reboot is required. To disable the hypervisor: svcadm disable svc:/milestone/xvm This will use bootadm(1M) to convert an existing menu.lst entry so that on next reboot Solaris will start on hardware. Please be aware that using this service will result in any tweaked parameters being discarded. In addition, the service will install a default value for the dom0_mem property (half of system memory or 2048MB, whichever value is lower). It is still possible to manually edit the menu.lst file, however use of the service will discard your changes. Gary
Hi Mark! Thanks for the info.> To enable the hypervisor: > > svcadm enable svc:/milestone/xvm > > This will use bootadm(1M) to convert an existing menu.lst entry so that > on next reboot the hypervisor will be started. The store service, > svc:/system/xvm/store, goes into maintenance to indicate that a reboot is > required. > > To disable the hypervisor: > > svcadm disable svc:/milestone/xvm > > This will use bootadm(1M) to convert an existing menu.lst entry so that > on next reboot Solaris will start on hardware. > > Please be aware that using this service will result in any tweaked parameters > being discarded. In addition, the service will install a default value for > the dom0_mem property (half of system memory or 2048MB, whichever value > is lower). > > It is still possible to manually edit the menu.lst file, however use of the > service will discard your changes.So do I understand correctly that if I ignore the new milestone I can still use the old menu.lst-based behavior? I like the existing setup much better. Now, I can decide if I want the hypervisor within my custom timeout _when booting_. Using the milestone, I would have to decide about the next boot _while shutting down_. Usually I boot my test systems for whatever I want to test _now_. Is this maybe a solution looking for a problem? Of course if the old behavior is still available I don''t really mind... Thanks -- Volker -- ------------------------------------------------------------------------ Volker A. Brandt Consulting and Support for Sun Solaris Brandt & Brandt Computer GmbH WWW: http://www.bb-c.de/ Am Wiesenpfad 6, 53340 Meckenheim Email: vab@bb-c.de Handelsregister: Amtsgericht Bonn, HRB 10513 Schuhgröße: 45 Geschäftsführer: Rainer J. H. Brandt und Volker A. Brandt
On Mon, Oct 12, 2009 at 02:37:09PM +0200, Volker A. Brandt wrote:> > It is still possible to manually edit the menu.lst file, however use of the > > service will discard your changes. > > So do I understand correctly that if I ignore the new milestone I can > still use the old menu.lst-based behavior? I like the existing setup > much better. Now, I can decide if I want the hypervisor within my custom > timeout _when booting_.Correct.> Using the milestone, I would have to decide about the next boot _while > shutting down_. Usually I boot my test systems for whatever I want > to test _now_.Two points: 1) the best way to do what you want is create a separate BE with beadm, then convert that to xVM. 2) We believe that most machines outside of test environments will remain in one or the other of the two environments. For other machines, the BE method is still best, as it allows a separate update schedule for when you are doing hypervisor test stuff regards john
> > Using the milestone, I would have to decide about the next boot _while > > shutting down_. Usually I boot my test systems for whatever I want > > to test _now_. > > Two points: > > 1) the best way to do what you want is create a separate BE with beadm, > then convert that to xVM. > > 2) We believe that most machines outside of test environments will > remain in one or the other of the two environments. For other machines, > the BE method is still best, as it allows a separate update schedule for > when you are doing hypervisor test stuffOK, I''ll certainly try it when it shows up in the build. I hadn''t considered a BE, I''ve just been picking a mode from the GRUB menu. Regards -- Volker -- ------------------------------------------------------------------------ Volker A. Brandt Consulting and Support for Sun Solaris Brandt & Brandt Computer GmbH WWW: http://www.bb-c.de/ Am Wiesenpfad 6, 53340 Meckenheim Email: vab@bb-c.de Handelsregister: Amtsgericht Bonn, HRB 10513 Schuhgröße: 45 Geschäftsführer: Rainer J. H. Brandt und Volker A. Brandt