Re: [gentoo-user] Anyone running a gentoo guest on virtualbox?

2011-06-01 Thread Joost Roeleveld
On Wednesday 01 June 2011 11:52:25 Pandu Poluan wrote:
 On Wed, Jun 1, 2011 at 03:35, Mick michaelkintz...@gmail.com wrote:
  On Tuesday 31 May 2011 21:02:46 Alan McKinnon wrote:
  I see.  In my head it is as if we're going against the udev principle of
  populating required device nodes.  If udev does not start, isn't it time
  to head for the nearest LiveCD, or must we ensure that every breakage
  is fixable in single-user mode?
 
 There are cases for each, but I personally prefer going single-user.
 Especially when working on virtualized servers.

+1
Even though with virtualized servers, using Xen, it's possible to access the 
filesystem easily from the host.

I do prefer to have the option for a single-usermode to fix things as I tend to 
disconnect CD-drives in servers to keep the CD-drives from killing themselves 
by being powered non-stop.

--
Joost



Re: [gentoo-user] Anyone running a gentoo guest on virtualbox?

2011-06-01 Thread Alan McKinnon
Apparently, though unproven, at 22:35 on Tuesday 31 May 2011, Mick did opine 
thusly:

  Considering that ~250 devices consumes a teeny-weeny bit of disk space
  and they are hidden from view normally, I say it's worth it leaving them
  in. Which is what vapier also says.
 
 I see.  In my head it is as if we're going against the udev principle of 
 populating required device nodes.  If udev does not start, isn't it time
 to  head for the nearest LiveCD, or must we ensure that every breakage is
 fixable in single-user mode?

I don't think we have to go that far. I think we should at least take 
reasonable steps to ensure the single-user mode works at all.


-- 
alan dot mckinnon at gmail dot com



Re: [gentoo-user] Anyone running a gentoo guest on virtualbox?

2011-06-01 Thread Pandu Poluan
On Wed, Jun 1, 2011 at 15:10, Joost Roeleveld jo...@antarean.org wrote:
 On Wednesday 01 June 2011 11:52:25 Pandu Poluan wrote:
 On Wed, Jun 1, 2011 at 03:35, Mick michaelkintz...@gmail.com wrote:
  On Tuesday 31 May 2011 21:02:46 Alan McKinnon wrote:
  I see.  In my head it is as if we're going against the udev principle of
  populating required device nodes.  If udev does not start, isn't it time
  to head for the nearest LiveCD, or must we ensure that every breakage
  is fixable in single-user mode?

 There are cases for each, but I personally prefer going single-user.
 Especially when working on virtualized servers.

 +1
 Even though with virtualized servers, using Xen, it's possible to access the
 filesystem easily from the host.


Well, in my case, the servers ran as VMs on top of VMware in my Cloud
Provider's infrastructure...

No consoling into the hypervisor, understandably. And even worse:
Can't attach the virtual hard disk to another live VM.

Either I have to boot with a LiveCD, or go into single-user. The
former took quite some time to boot, so I really prefer the latter.

Rgds,
-- 
Pandu E Poluan
~ IT Optimizer ~
Visit my Blog: http://pepoluan.posterous.com



Re: [gentoo-user] Anyone running a gentoo guest on virtualbox?

2011-05-31 Thread Alan McKinnon
On Tue, May 31, 2011 at 3:14 AM, Pandu Poluan pa...@poluan.info wrote:
 Meh, I clicked 'Send' too fast.

 *My* suggested solution:

 Generate an initramfs containing udev. The hands-down easiest way is
 using genkernel's 'only create an initramfs' switch (sorry I forgot
 what exactly).

good god no, please, anything but genkernel.

That thing is an attempt to emulate binary distros which require an
initramfs to work properly (for any sane definition of work) as the
person building the installer has no idea what hardware the user will
have. In Gentoo the user knows exactly what they have so there's no
need for a gigantic hardware-detecting workaround at boot time.

 This needs to be done exactly once throughout the life of your VM.

 (To the herd of Gentoo graybeards, feel free to CMIIW)

Or wait a few days for vapier's (posting under his other name of
spanky) sane advice to be implemented. His proposal is the sole voice
of reason in that bug thread



 Another alternative would be to mknod all required devices for
 booting. But, as evidenced in the bug I've linked to earlier, you
 might have to create more than 20 devs. Not a good use of time, if you
 ask me. Except if you're one of the guys doing the bug exorcising :)

 Oh, and please forgive my top-postings. Gmail's Java mobile client sucks.

 Rgds,


 On 2011-05-31, walt w41...@gmail.com wrote:
 In preparation for the upcoming upgrade to gnome3, I've installed
 the latest gentoo snapshot to a new virtualbox machine.  (So I can
 trash my virtual gentoo machine instead of my real gentoo machine :)

 The virtual install went perfectly AFAICT, except for building a new
 customized kernel for the gentoo virtualbox machine.

 Here's what I did to configure my new customized gentoo kernel:

 I booted the gentoo install iso image in virtualbox and did lspci -k
 and wrote down all the drivers it displayed.

 I also booted my virtualbox ubuntu machine and did lspci -k and again
 wrote down all the listed drivers.  (Only one extra driver showed up
 in ubuntu and I included it in my list of drivers to-be-installed.)

 I configured my new gentoo custom kernel to use all of the drivers I'd
 gathered from the steps above, and compiled and installed it without
 any problems.

 However, when I reboot the virtual gentoo guest machine with my new
 customized kernel, the boot hangs forever after discovering devices
 and mounting the root partition.ro.

 Obviously I've configured my custom kernel incorrectly, but how?

 If any of you have virtualbox guest gentoo machines running with a
 custom kernel, would you please post your guest .config file for my
 edification?

 Many thanks!





 --
 --
 Pandu E Poluan - IT Optimizer
 My website: http://pandu.poluan.info/





-- 
Alan McKinnon
alan dot mckinnon at gmail dot com



Re: [gentoo-user] Anyone running a gentoo guest on virtualbox?

2011-05-31 Thread Pandu Poluan
On Tue, May 31, 2011 at 13:56, Alan McKinnon alan.mckin...@gmail.com wrote:
 On Tue, May 31, 2011 at 3:14 AM, Pandu Poluan pa...@poluan.info wrote:
 Meh, I clicked 'Send' too fast.

 *My* suggested solution:

 Generate an initramfs containing udev. The hands-down easiest way is
 using genkernel's 'only create an initramfs' switch (sorry I forgot
 what exactly).

 good god no, please, anything but genkernel.

 That thing is an attempt to emulate binary distros which require an
 initramfs to work properly (for any sane definition of work) as the
 person building the installer has no idea what hardware the user will
 have. In Gentoo the user knows exactly what they have so there's no
 need for a gigantic hardware-detecting workaround at boot time.

 This needs to be done exactly once throughout the life of your VM.

 (To the herd of Gentoo graybeards, feel free to CMIIW)

 Or wait a few days for vapier's (posting under his other name of
 spanky) sane advice to be implemented. His proposal is the sole voice
 of reason in that bug thread


True. But I was having problem installing 2 servers on top of XenServer.

So I cheated and ran 'genkernel initramfs' exactly once. At least I
got myself a booting system. :-)

When SpanKY's makedev gets stabilized and pushed to baselayout, I'll
then happily ditch the genkernel cheat for my next VMs :-)

Rgds,
-- 
Pandu E Poluan
~ IT Optimizer ~
Visit my Blog: http://pepoluan.posterous.com



Re: [gentoo-user] Anyone running a gentoo guest on virtualbox?

2011-05-31 Thread Mick
On Tuesday 31 May 2011 08:07:24 Pandu Poluan wrote:
 On Tue, May 31, 2011 at 13:56, Alan McKinnon alan.mckin...@gmail.com 
wrote:
  On Tue, May 31, 2011 at 3:14 AM, Pandu Poluan pa...@poluan.info wrote:
  Meh, I clicked 'Send' too fast.
  
  *My* suggested solution:
  
  Generate an initramfs containing udev. The hands-down easiest way is
  using genkernel's 'only create an initramfs' switch (sorry I forgot
  what exactly).
  
  good god no, please, anything but genkernel.
  
  That thing is an attempt to emulate binary distros which require an
  initramfs to work properly (for any sane definition of work) as the
  person building the installer has no idea what hardware the user will
  have. In Gentoo the user knows exactly what they have so there's no
  need for a gigantic hardware-detecting workaround at boot time.
  
  This needs to be done exactly once throughout the life of your VM.
  
  (To the herd of Gentoo graybeards, feel free to CMIIW)
  
  Or wait a few days for vapier's (posting under his other name of
  spanky) sane advice to be implemented. His proposal is the sole voice
  of reason in that bug thread
 
 True. But I was having problem installing 2 servers on top of XenServer.
 
 So I cheated and ran 'genkernel initramfs' exactly once. At least I
 got myself a booting system. :-)
 
 When SpanKY's makedev gets stabilized and pushed to baselayout, I'll
 then happily ditch the genkernel cheat for my next VMs :-)

Are you sure that manually creating /dev/console and /dev/null isn't all that 
is required?  The rest of the devices will be created by udev when it runs at 
boot time.

-- 
Regards,
Mick


signature.asc
Description: This is a digitally signed message part.


Re: [gentoo-user] Anyone running a gentoo guest on virtualbox?

2011-05-31 Thread Alan McKinnon
Apparently, though unproven, at 21:27 on Tuesday 31 May 2011, Mick did opine 
thusly:

 On Tuesday 31 May 2011 08:07:24 Pandu Poluan wrote:
  On Tue, May 31, 2011 at 13:56, Alan McKinnon alan.mckin...@gmail.com
 
 wrote:
   On Tue, May 31, 2011 at 3:14 AM, Pandu Poluan pa...@poluan.info wrote:
   Meh, I clicked 'Send' too fast.
   
   *My* suggested solution:
   
   Generate an initramfs containing udev. The hands-down easiest way is
   using genkernel's 'only create an initramfs' switch (sorry I forgot
   what exactly).
   
   good god no, please, anything but genkernel.
   
   That thing is an attempt to emulate binary distros which require an
   initramfs to work properly (for any sane definition of work) as the
   person building the installer has no idea what hardware the user will
   have. In Gentoo the user knows exactly what they have so there's no
   need for a gigantic hardware-detecting workaround at boot time.
   
   This needs to be done exactly once throughout the life of your VM.
   
   (To the herd of Gentoo graybeards, feel free to CMIIW)
   
   Or wait a few days for vapier's (posting under his other name of
   spanky) sane advice to be implemented. His proposal is the sole voice
   of reason in that bug thread
  
  True. But I was having problem installing 2 servers on top of XenServer.
  
  So I cheated and ran 'genkernel initramfs' exactly once. At least I
  got myself a booting system. :-)
  
  When SpanKY's makedev gets stabilized and pushed to baselayout, I'll
  then happily ditch the genkernel cheat for my next VMs :-)
 
 Are you sure that manually creating /dev/console and /dev/null isn't all
 that is required?  The rest of the devices will be created by udev when it
 runs at boot time.

null and console are the absolute irreducible minimum but there's one that can 
be dispensed with if the correct kernel option is enabled.

We don't need everything that makedev traditionally provided (like every block 
device type known to man, floppys and ancient ptys) but the rest number about 
~250 and are useful in single-user mode if udev fails to start.

Considering that ~250 devices consumes a teeny-weeny bit of disk space and 
they are hidden from view normally, I say it's worth it leaving them in. Which 
is what vapier also says.


-- 
alan dot mckinnon at gmail dot com



Re: [gentoo-user] Anyone running a gentoo guest on virtualbox?

2011-05-31 Thread Dale

Alan McKinnon wrote:


Considering that ~250 devices consumes a teeny-weeny bit of disk space and
they are hidden from view normally, I say it's worth it leaving them in. Which
is what vapier also says.

   


+1  They are tiny plus when devfs mounts, they aren't visible anymore if 
I recall correctly.  Doesn't devfs mount on top of them?


Dale

:-)  :-)



Re: [gentoo-user] Anyone running a gentoo guest on virtualbox?

2011-05-31 Thread Alan McKinnon
Apparently, though unproven, at 22:12 on Tuesday 31 May 2011, Dale did opine 
thusly:

 Alan McKinnon wrote:
  Considering that ~250 devices consumes a teeny-weeny bit of disk space
  and they are hidden from view normally, I say it's worth it leaving them
  in. Which is what vapier also says.
 
 +1  They are tiny plus when devfs mounts, they aren't visible anymore if
 I recall correctly.  Doesn't devfs mount on top of them?

Well that's what hidden from view normally evaluates to.

But it's not devfs - that was an abomination that should never have been 
suffered to live. It's mere existence offended GregKH so much that he whipped 
up the beginnings of udev so that he might never see devfs ever again

It's udev and is normally mounted on a tmpfs

-- 
alan dot mckinnon at gmail dot com



Re: [gentoo-user] Anyone running a gentoo guest on virtualbox?

2011-05-31 Thread Mick
On Tuesday 31 May 2011 21:02:46 Alan McKinnon wrote:
 Apparently, though unproven, at 21:27 on Tuesday 31 May 2011, Mick did
 opine
 
 thusly:
  On Tuesday 31 May 2011 08:07:24 Pandu Poluan wrote:
   On Tue, May 31, 2011 at 13:56, Alan McKinnon alan.mckin...@gmail.com
  
  wrote:
On Tue, May 31, 2011 at 3:14 AM, Pandu Poluan pa...@poluan.info 
wrote:
Meh, I clicked 'Send' too fast.

*My* suggested solution:

Generate an initramfs containing udev. The hands-down easiest way is
using genkernel's 'only create an initramfs' switch (sorry I forgot
what exactly).

good god no, please, anything but genkernel.

That thing is an attempt to emulate binary distros which require an
initramfs to work properly (for any sane definition of work) as the
person building the installer has no idea what hardware the user will
have. In Gentoo the user knows exactly what they have so there's no
need for a gigantic hardware-detecting workaround at boot time.

This needs to be done exactly once throughout the life of your VM.

(To the herd of Gentoo graybeards, feel free to CMIIW)

Or wait a few days for vapier's (posting under his other name of
spanky) sane advice to be implemented. His proposal is the sole voice
of reason in that bug thread
   
   True. But I was having problem installing 2 servers on top of
   XenServer.
   
   So I cheated and ran 'genkernel initramfs' exactly once. At least I
   got myself a booting system. :-)
   
   When SpanKY's makedev gets stabilized and pushed to baselayout, I'll
   then happily ditch the genkernel cheat for my next VMs :-)
  
  Are you sure that manually creating /dev/console and /dev/null isn't all
  that is required?  The rest of the devices will be created by udev when
  it runs at boot time.
 
 null and console are the absolute irreducible minimum but there's one that
 can be dispensed with if the correct kernel option is enabled.
 
 We don't need everything that makedev traditionally provided (like every
 block device type known to man, floppys and ancient ptys) but the rest
 number about ~250 and are useful in single-user mode if udev fails to
 start.
 
 Considering that ~250 devices consumes a teeny-weeny bit of disk space and
 they are hidden from view normally, I say it's worth it leaving them in.
 Which is what vapier also says.

I see.  In my head it is as if we're going against the udev principle of 
populating required device nodes.  If udev does not start, isn't it time to 
head for the nearest LiveCD, or must we ensure that every breakage is fixable 
in single-user mode?
-- 
Regards,
Mick


signature.asc
Description: This is a digitally signed message part.


Re: [gentoo-user] Anyone running a gentoo guest on virtualbox?

2011-05-31 Thread Dale

Alan McKinnon wrote:

Apparently, though unproven, at 22:12 on Tuesday 31 May 2011, Dale did opine
thusly:

   

Alan McKinnon wrote:
 

Considering that ~250 devices consumes a teeny-weeny bit of disk space
and they are hidden from view normally, I say it's worth it leaving them
in. Which is what vapier also says.
   

+1  They are tiny plus when devfs mounts, they aren't visible anymore if
I recall correctly.  Doesn't devfs mount on top of them?
 

Well that's what hidden from view normally evaluates to.

But it's not devfs - that was an abomination that should never have been
suffered to live. It's mere existence offended GregKH so much that he whipped
up the beginnings of udev so that he might never see devfs ever again

It's udev and is normally mounted on a tmpfs

   


Correct.  I was thinking about the old way.  Still mounted on top of and 
hidden as you say.


Dale

:-)  :-)



Re: [gentoo-user] Anyone running a gentoo guest on virtualbox?

2011-05-31 Thread Pandu Poluan
On Wed, Jun 1, 2011 at 03:35, Mick michaelkintz...@gmail.com wrote:
 On Tuesday 31 May 2011 21:02:46 Alan McKinnon wrote:
 Apparently, though unproven, at 21:27 on Tuesday 31 May 2011, Mick did
 opine

 thusly:
  On Tuesday 31 May 2011 08:07:24 Pandu Poluan wrote:
   On Tue, May 31, 2011 at 13:56, Alan McKinnon alan.mckin...@gmail.com

- 8 - massive snippage - 8 -

   When SpanKY's makedev gets stabilized and pushed to baselayout, I'll
   then happily ditch the genkernel cheat for my next VMs :-)
 
  Are you sure that manually creating /dev/console and /dev/null isn't all
  that is required?  The rest of the devices will be created by udev when
  it runs at boot time.


Most probably so. But at that point, I was pressed for time. Had the
system need only /dev/{console,null} then all will be well. If not?
Then another cycle of LiveCD-mount-mknod-restart.

Much faster to just `genkernel initramfs` while waiting for the snafus
to be fixed

(Well, that, and I'm lazy)

 null and console are the absolute irreducible minimum but there's one that
 can be dispensed with if the correct kernel option is enabled.

 We don't need everything that makedev traditionally provided (like every
 block device type known to man, floppys and ancient ptys) but the rest
 number about ~250 and are useful in single-user mode if udev fails to
 start.

 Considering that ~250 devices consumes a teeny-weeny bit of disk space and
 they are hidden from view normally, I say it's worth it leaving them in.
 Which is what vapier also says.

Agree.

 I see.  In my head it is as if we're going against the udev principle of
 populating required device nodes.  If udev does not start, isn't it time to
 head for the nearest LiveCD, or must we ensure that every breakage is fixable
 in single-user mode?

There are cases for each, but I personally prefer going single-user.
Especially when working on virtualized servers.

Rgds,
-- 
Pandu E Poluan
~ IT Optimizer ~
Visit my Blog: http://pepoluan.posterous.com



[gentoo-user] Anyone running a gentoo guest on virtualbox?

2011-05-30 Thread walt
In preparation for the upcoming upgrade to gnome3, I've installed
the latest gentoo snapshot to a new virtualbox machine.  (So I can
trash my virtual gentoo machine instead of my real gentoo machine :)

The virtual install went perfectly AFAICT, except for building a new
customized kernel for the gentoo virtualbox machine.

Here's what I did to configure my new customized gentoo kernel:

I booted the gentoo install iso image in virtualbox and did lspci -k
and wrote down all the drivers it displayed.

I also booted my virtualbox ubuntu machine and did lspci -k and again
wrote down all the listed drivers.  (Only one extra driver showed up
in ubuntu and I included it in my list of drivers to-be-installed.)

I configured my new gentoo custom kernel to use all of the drivers I'd
gathered from the steps above, and compiled and installed it without
any problems.

However, when I reboot the virtual gentoo guest machine with my new
customized kernel, the boot hangs forever after discovering devices
and mounting the root partition.ro.

Obviously I've configured my custom kernel incorrectly, but how?

If any of you have virtualbox guest gentoo machines running with a
custom kernel, would you please post your guest .config file for my
edification?

Many thanks!




Re: [gentoo-user] Anyone running a gentoo guest on virtualbox?

2011-05-30 Thread Pandu Poluan
Are you using a recent stage3 tarball? If so, I suspect your booting
problem has got something to do with this bug:

http://bugs.gentoo.org/show_bug.cgi?id=368597

Rgds,


On 2011-05-31, walt w41...@gmail.com wrote:
 In preparation for the upcoming upgrade to gnome3, I've installed
 the latest gentoo snapshot to a new virtualbox machine.  (So I can
 trash my virtual gentoo machine instead of my real gentoo machine :)

 The virtual install went perfectly AFAICT, except for building a new
 customized kernel for the gentoo virtualbox machine.

 Here's what I did to configure my new customized gentoo kernel:

 I booted the gentoo install iso image in virtualbox and did lspci -k
 and wrote down all the drivers it displayed.

 I also booted my virtualbox ubuntu machine and did lspci -k and again
 wrote down all the listed drivers.  (Only one extra driver showed up
 in ubuntu and I included it in my list of drivers to-be-installed.)

 I configured my new gentoo custom kernel to use all of the drivers I'd
 gathered from the steps above, and compiled and installed it without
 any problems.

 However, when I reboot the virtual gentoo guest machine with my new
 customized kernel, the boot hangs forever after discovering devices
 and mounting the root partition.ro.

 Obviously I've configured my custom kernel incorrectly, but how?

 If any of you have virtualbox guest gentoo machines running with a
 custom kernel, would you please post your guest .config file for my
 edification?

 Many thanks!





-- 
--
Pandu E Poluan - IT Optimizer
My website: http://pandu.poluan.info/



Re: [gentoo-user] Anyone running a gentoo guest on virtualbox?

2011-05-30 Thread Pandu Poluan
Meh, I clicked 'Send' too fast.

*My* suggested solution:

Generate an initramfs containing udev. The hands-down easiest way is
using genkernel's 'only create an initramfs' switch (sorry I forgot
what exactly).

This needs to be done exactly once throughout the life of your VM.

(To the herd of Gentoo graybeards, feel free to CMIIW)

Another alternative would be to mknod all required devices for
booting. But, as evidenced in the bug I've linked to earlier, you
might have to create more than 20 devs. Not a good use of time, if you
ask me. Except if you're one of the guys doing the bug exorcising :)

Oh, and please forgive my top-postings. Gmail's Java mobile client sucks.

Rgds,


On 2011-05-31, walt w41...@gmail.com wrote:
 In preparation for the upcoming upgrade to gnome3, I've installed
 the latest gentoo snapshot to a new virtualbox machine.  (So I can
 trash my virtual gentoo machine instead of my real gentoo machine :)

 The virtual install went perfectly AFAICT, except for building a new
 customized kernel for the gentoo virtualbox machine.

 Here's what I did to configure my new customized gentoo kernel:

 I booted the gentoo install iso image in virtualbox and did lspci -k
 and wrote down all the drivers it displayed.

 I also booted my virtualbox ubuntu machine and did lspci -k and again
 wrote down all the listed drivers.  (Only one extra driver showed up
 in ubuntu and I included it in my list of drivers to-be-installed.)

 I configured my new gentoo custom kernel to use all of the drivers I'd
 gathered from the steps above, and compiled and installed it without
 any problems.

 However, when I reboot the virtual gentoo guest machine with my new
 customized kernel, the boot hangs forever after discovering devices
 and mounting the root partition.ro.

 Obviously I've configured my custom kernel incorrectly, but how?

 If any of you have virtualbox guest gentoo machines running with a
 custom kernel, would you please post your guest .config file for my
 edification?

 Many thanks!





-- 
--
Pandu E Poluan - IT Optimizer
My website: http://pandu.poluan.info/