Ron,
Solaris was not reliable 15 years ago and it is not reliable even today.
Drivers and handling drivers have not been tackled at any level.

I would agree with you or anyone who says a reboot should be a part of
handling the new encapsulated disks for mounted partitions on a disk.

-GGR


On Mon, Mar 3, 2008 at 2:34 AM, Ronald S Karr <[EMAIL PROTECTED]> wrote:

> When I first wrote the script (many, many, many moons ago), Solaris wasn't
> reliable about getting the disk driver to update its view of the partition
> table, which would cause confusion in some cases, so when I wrote this stuff
> I was overly conservative.  But, even now, it is a bit tricky trying to
> prove that none of the partitions are in use, particularly from a shell
> script (and vxencap is a shell script).
>
> Fixing the script to avoid reboots has been discussed quite a few times,
> but nobody has wanted to prioritize tackling a script like this that
> developers have been updating for over 15 years.
> --     tron |-<=>-|
>
> On Mar 2, 2008, at 10:25 AM, Asim Zuberi wrote:
>
> Hi Rajiv / Ron --
>
> Thanks for your responses and your assistance in explaining me the
> concepts of reboot required for encapsulation.
>
> Rajiv -- Actually, I've tried your proposed suggestion already. I used a
> non-root-disk to encapsulate by using
> "/etc/vx/bin/vxencap" command. The command ran clean, but it didn't
> encapsulate the root drive. I couldn't boot the
> system to complete the encapsulation process.
>
> As Ron has pointed out in his email the script /etc/init.d/vxvm-reconfig
> handles the encapsulation, which does reboot
> the system. Maybe it is time for VRTS developers to modify the
> encapsulation process without a reboot so that this
> can be achieved on the fly.
>
> The reason I am enquiring the details of the encapsulation process is
> because of capturing Solaris Zones under VxVM
> control. To make Solaris Zones part of VxVM every time a zone is created,
> making it impossible to reboot the physical
> system so often. So we're looking into options where a reboot of the
> physical can be avoided.
>
> Question is has someone already identified a process to accomplish this
> task? I'd be happy to try out the trials and share
> the workarounds if I get some pointers from VRTS engineers on this mailing
> lists.
>
>
>
> thanks!
> --Asim;
>
>
>
>
>  ------------------------------
> *From:* Rajiv Gunja [mailto:[EMAIL PROTECTED]<[EMAIL PROTECTED]>]
>
> *Sent:* Friday, February 29, 2008 7:10 PM
> *To:* Ronald S Karr; Asim Zuberi; [EMAIL PROTECTED]
> *Subject:* Re: [Veritas-vx] encapsulation requires a reboot - why?
>
> Asim,
>
> I am no Veritas expert, but speaking only from logic, here is how I see
> fit to answer you:
>
>
>    1. Encapsulation of a disk is done on disks which are NOT under
>    Veritas control, but is in the system.
>    2. Encapsulating a disk on the system requires a reboot ONLY if the
>    said disk has current slices mounted and/or has applications running on it.
>    3. Encapsulation/Veritas take over of a disk usually done for disks
>    which are single and you need a way to expand or create software mirror of
>    the disk.
>
>
> Hope this sheds some light on your question. So the next test you need to
> do is to encapsulate a disk which has no mounted slices and see if it
> requires a reboot.
>
> -GGR
>
> On Fri, Feb 29, 2008 at 4:59 PM, Ronald S Karr <[EMAIL PROTECTED]> wrote:
>
> > Encapsulation is done by the script /etc/init.d/vxvm-reconfig.  This
> > handles a variety of things, some of which can only be handled during
> > a reboot, when everything is clean.  If none of those things are going
> > on, you could run /etc/init.d/vxvm-reconfig.  But, this script will
> > reboot your system automatically if you encapsulate a volume that is
> > in use, so be careful.  Don't expect any sympathy from support if this
> > doesn't go well.
> > --
> >     Ronald S. Karr
> >     tron |-<=>-|    [EMAIL PROTECTED]
> >
> > On Feb 29, 2008, at 10:06 AM, Asim Zuberi wrote:
> > > Thanks to the following for their quick responses
> > >
> > > Pat Owens
> > > Darren Dunham
> > >
> > > Follow-up question: If I am encapsulating a non-root disk, such as a
> > > data
> > > disk, it appears I still need to reboot the entire
> > > physical system. Is there a work around to avoid a reboot of the
> > > physical
> > > system, if only a data disk is being encapsulated?
> > >
> > > For testing purposes, I attempted to encapsulate a slice of a non-
> > > root disk
> > > on a system by using /etc/vx/bin/vxencap command.
> > > The command ran clean but the encapsulation didn't take place until
> > > a reboot
> > > occurs.
> > >
> > > Please advise.
> > >
> > > thanks!
> > > --Asim;
> > >
> > >
> > >
> > >
> > > =]  -----Original Message-----
> > > =]  From: [EMAIL PROTECTED]
> > > =]  [mailto:[EMAIL PROTECTED] On
> > > =]  Behalf Of Asim Zuberi
> > > =]  Sent: Friday, February 29, 2008 8:48 AM
> > > =]  To: [EMAIL PROTECTED]
> > > =]  Subject: [Veritas-vx] encapsulation requires a reboot - why?
> > > =]
> > > =]
> > > =]
> > > =]  hi All --
> > > =]
> > > =]  Why the process of encapsulation (of a disk) requires a
> > > =]  reboot of the system?
> > > =]  Is there a way, it can be avoided?
> > > =]
> > > =]  thanks!
> > > =]  --Asim;
> > > =]
> > > =]  _______________________________________________
> > > =]  Veritas-vx maillist  -  Veritas-vx@mailman.eng.auburn.edu
> > > =]  http://mailman.eng.auburn.edu/mailman/listinfo/veritas-vx
> > > =]
> > >
> > > _______________________________________________
> > > Veritas-vx maillist  -  Veritas-vx@mailman.eng.auburn.edu
> > > http://mailman.eng.auburn.edu/mailman/listinfo/veritas-vx
> >
> > _______________________________________________
> >  Veritas-vx maillist  -  Veritas-vx@mailman.eng.auburn.edu
> > http://mailman.eng.auburn.edu/mailman/listinfo/veritas-vx
> >
>
>
>     Ronald S. Karr
>     tron |-<=>-|    [EMAIL PROTECTED]
>
>
>
>
_______________________________________________
Veritas-vx maillist  -  Veritas-vx@mailman.eng.auburn.edu
http://mailman.eng.auburn.edu/mailman/listinfo/veritas-vx

Reply via email to