zfs discuss - Jan 2007 - can I use zfs on just a partition?

Just want to verify, if I have say, 1 160GB disk, can I format it so that the
first say 40GB is my main UFS parition with the base OS install, and then make
the rest of the disk zfs?  Or even better yet, for testing purposes make two
60GB partitions out of the rest of it and make them a *mirror*?  Or do I have to
give zfs the entire disk?
 
 
This message posted from opensolaris.org

On 1/25/07, Tim Cook <tcook at datalink.com>
wrote:
> Just want to verify, if I have say, 1 160GB disk, can I format it so that
the first say 40GB is my main UFS parition with the base OS install, and then
make the rest of the disk zfs?  Or even better yet, for testing purposes make
two 60GB partitions out of the rest of it and make them a *mirror*?  Or do I
have to give zfs the entire disk?
If you''re talking about slices, yes. (partitions). You can give zfs
just a slice of the disk, while giving UFS another slice. You can
mirror across slices (though performance will be sub optimal).

Take note though, that giving zfs the entire disk gives a possible
performance win, as zfs will only enable the write cache for the disk
if it is given the entire disk.

A caveat: giving the whole disk to a zfs zpool results in an EFI label
being assigned: some motherboard BIOSes refuse to boot with a hard
disk that has an EFI label.
-- 
Regards,
Jeremy

On 25/1/07 3:16, "Jeremy Teo" <white.wristband at gmail.com>
wrote:
> On 1/25/07, Tim Cook <tcook at datalink.com> wrote:
>> Just want to verify, if I have say, 1 160GB disk, can I format it so
that the
>> first say 40GB is my main UFS parition with the base OS install, and
then
>> make the rest of the disk zfs?  Or even better yet, for testing
purposes make
>> two 60GB partitions out of the rest of it and make them a *mirror*?  Or
do I
>> have to give zfs the entire disk?
> 
> If you''re talking about slices, yes. (partitions). You can give
zfs
> just a slice of the disk, while giving UFS another slice. You can
> mirror across slices (though performance will be sub optimal).
The zpool man page suggests that you can use a regular file as a backing
store, though this is only really intended for experimental purposes.

Cheers,

Chris

Well, the system can only have one disk, so giving it the full disk
isn''t really an option unless they''ve finally gotten the whole
boot from a zfs disk figured out.
 
 
This message posted from opensolaris.org

Tim Cook wrote:
> Just want to verify, if I have say, 1 160GB disk, can I format it so that
the first say 40GB is my main UFS parition with the base OS install, and then
make the rest of the disk zfs?  Or even better yet, for testing purposes make
two 60GB partitions out of the rest of it and make them a *mirror*?  Or do I
have to give zfs the entire disk?
yes you can do that and many of us do exactly that on our laptops.

I have even done the two slices in the VTOC as sides of the ZFS mirror.

Though on a 4200rpm disk that is REALLY bad for performance of writes :-)
-- 
Darren J Moffat

I guess I should clarify what I''m doing.

Essentially I''d like to have the / and swap on the first 60GB of the
disk.  Then use the remaining 100GB as a zfs partition to setup zones on. 
Obviously the snapshots are extremely useful in such a setup :)

Does my plan sound feasible from both a usability and performance standpoint?
 
 
This message posted from opensolaris.org

On 1/25/07, Tim Cook <tcook at datalink.com>
wrote:
> I guess I should clarify what I''m doing.
>
> Essentially I''d like to have the / and swap on the first 60GB of
the disk.  Then use the remaining 100GB as a zfs partition to setup zones on. 
Obviously the snapshots are extremely useful in such a setup :)
>
> Does my plan sound feasible from both a usability and performance
standpoint?
Yes, however 60GB for root/swap is probably overkill.  My notebook''s
setup:

s0: 10GB -> root
s1: 2GB -> swap
s3: 10GB -> root
s4: <the rest> -> ZFS

(I use s0 and s3 for Live Upgrade)

-- 
Eric Enright

Hi Tim,
> Essentially I''d like to have the / and swap on the first 60GB of
the disk.  Then use the remaining 100GB as a zfs partition to setup zones on. 
Obviously the snapshots are extremely useful in such a setup :)
> 
> Does my plan sound feasible from both a usability and performance
standpoint?
yes, it works, I do it on my laptop all the time:

# format
Searching for disks...done


AVAILABLE DISK SELECTIONS:
       0. c0d0 <DEFAULT cyl 48451 alt 2 hd 64 sec 63>
          /pci at 0,0/pci-ide at 14,1/ide at 0/cmdk at 0,0
Specify disk (enter its number): 0
selecting c0d0
Controller working list found
[disk formatted, defect list found]
Warning: Current Disk has mounted partitions.
/dev/dsk/c0d0s0 is currently mounted on /. Please see umount(1M).
/dev/dsk/c0d0s1 is currently used by swap. Please see swap(1M).
/dev/dsk/c0d0s3 is part of active ZFS pool poolchen. Please see zpool(1M).
/dev/dsk/c0d0s4 is in use for live upgrade /. Please see ludelete(1M).

c0d0s5 is also free and can be used as a third live upgrade partition.

My recommendation: Use at least 2 slices for the OS so you can enjoy live
upgrade, one for swap and the rest for ZFS.

Performance-wise, this is of course not optimal, but perfectly feasible. I have
an Acer Ferrari 4000 which is known to have a slow disk, but it still works
great for what I do (email, web, Solaris demos, presentations, occasional
video).

More complicated things are possible as well. The following blog entry:

  http://blogs.sun.com/solarium/entry/tetris_spielen_mit_zfs

(sorry it''s german) ilustrates how my 4 disks at home are sliced in
order
to get OS partitions on multiple disks, Swap and as much ZFS space as
possible at acceptable redundancy despite differently-sized disks. Check out
the graphic in the above entry to see what I mean. Works great (but I had to
use -f to zpool create :) ) and gives me enough performance for all my
home-serving needs.

Hope this helps,
   Constantin

-- 
Constantin Gonzalez                            Sun Microsystems GmbH, Germany
Platform Technology Group, Client Solutions                http://www.sun.de/
Tel.: +49 89/4 60 08-25 91                   http://blogs.sun.com/constantin/

Tim Cook wrote:
> I guess I should clarify what I''m doing.
> 
> Essentially I''d like to have the / and swap on the first 60GB of
the disk.  Then use the remaining 100GB as a zfs partition to setup zones on. 
Obviously the snapshots are extremely useful in such a setup :)
> 
> Does my plan sound feasible from both a usability and performance
standpoint?
>  
That''s exactly how I''m running my ferrari laptop.  Works
like a charm.

- Bart


-- 
Bart Smaalders			Solaris Kernel Performance
barts at cyber.eng.sun.com		http://blogs.sun.com/barts

Tim Cook wrote:
> I guess I should clarify what I''m doing.
> 
> Essentially I''d like to have the / and swap on the first 60GB of
the disk.  Then use the remaining 100GB as a zfs partition to setup zones on. 
Obviously the snapshots are extremely useful in such a setup :)
> 
> Does my plan sound feasible from both a usability and performance
standpoint?
That is exactly what I do on my laptop, with one exception.  For me it 
isn''t just my zones on ZFS but also my home directory and my builds
area
(using zfs clones to quickly get a new mercurial repository is great!).

-- 
Darren J Moffat

On Thu, Jan 25, 2007 at 05:46:09PM +0000, Darren J Moffat
wrote:
> >
> >Does my plan sound feasible from both a usability and performance 
> >standpoint?
> 
> That is exactly what I do on my laptop, with one exception.  For me it 
> isn''t just my zones on ZFS but also my home directory and my
builds area
> (using zfs clones to quickly get a new mercurial repository is great!).
for me, /usr/local is on the ZFS portion of my desktop''s disk.  I
always do a
snapshot before installing new software to it.  It''s *great*.  ;)

-brian
-- 
"The reason I don''t use Gnome: every single other window manager I
know of is
very powerfully extensible, where you can switch actions to different mouse
buttons. Guess which one is not, because it might confuse the poor users?
Here''s a hint: it''s not the small and fast one."       
--Linus

On Thu, 25 Jan 2007, Darren J Moffat wrote:
> Tim Cook wrote:
> > I guess I should clarify what I''m doing.
> >
> > Essentially I''d like to have the / and swap on the first 60GB
of the
> > disk.  Then use the remaining 100GB as a zfs partition to setup zones
> > on.  Obviously the snapshots are extremely useful in such a setup :)
> >
> > Does my plan sound feasible from both a usability and performance
standpoint?
For a laptop scenario, where you don''t have any option to add a 2nd
drive,
I agree with Darren and others.  But in any other usage scenario, I
don''t
think its a good idea.  Perhaps you could post some details about your
hardware environment and why its limited to a single disk drive?

You understand that if ZFS detects any data corruption and you don''t
have
any designed-in data redundancy, ZFS will make the pool unavailable and
you are SOL.
> That is exactly what I do on my laptop, with one exception.  For me it
> isn''t just my zones on ZFS but also my home directory and my
builds area
> (using zfs clones to quickly get a new mercurial repository is great!).
>
> --
Regards,

Al Hopper  Logical Approach Inc, Plano, TX.  al at logical-approach.com
           Voice: 972.379.2133 Fax: 972.379.2134  Timezone: US CDT
OpenSolaris.Org Community Advisory Board (CAB) Member - Apr 2005
             OpenSolaris Governing Board (OGB) Member - Feb 2006

When you do the initial install, how do you do the slicing?

Just create like:
/ 10G
swap 2G
/altroot 10G
/zfs restofdisk

Or do you just create the first three slices and leave the rest of the disk
untouched?  I understand the concept at this point, just trying to explain to a
third party exactly what they need to do to prep the system disk for me :)
 
 
This message posted from opensolaris.org

It''s a third party host, and I''ve been informed the cases they
use only
have room available for one hard drive.  It''s definitely not my first
choice, but it''s the only option I have at this point.

Tim Cook

-----Original Message-----
From: Al Hopper [mailto:al at logical-approach.com] 
Sent: Thursday, January 25, 2007 11:53 AM
To: Tim Cook
Cc: zfs-discuss at opensolaris.org
Subject: Re: [zfs-discuss] Re: can I use zfs on just a partition?

On Thu, 25 Jan 2007, Darren J Moffat wrote:
> Tim Cook wrote:
> > I guess I should clarify what I''m doing.
> >
> > Essentially I''d like to have the / and swap on the first 60GB
of the
> > disk.  Then use the remaining 100GB as a zfs partition to setup
zones
> > on.  Obviously the snapshots are extremely useful in such a setup :)
> >
> > Does my plan sound feasible from both a usability and performance
standpoint?

For a laptop scenario, where you don''t have any option to add a 2nd
drive,
I agree with Darren and others.  But in any other usage scenario, I
don''t
think its a good idea.  Perhaps you could post some details about your
hardware environment and why its limited to a single disk drive?

You understand that if ZFS detects any data corruption and you don''t
have
any designed-in data redundancy, ZFS will make the pool unavailable and
you are SOL.
> That is exactly what I do on my laptop, with one exception.  For me it
> isn''t just my zones on ZFS but also my home directory and my
builds
area
> (using zfs clones to quickly get a new mercurial repository is
great!).
>
> --
Regards,

Al Hopper  Logical Approach Inc, Plano, TX.  al at logical-approach.com
           Voice: 972.379.2133 Fax: 972.379.2134  Timezone: US CDT
OpenSolaris.Org Community Advisory Board (CAB) Member - Apr 2005
             OpenSolaris Governing Board (OGB) Member - Feb 2006

Hello Tim,

Thursday, January 25, 2007, 4:44:34 PM, you wrote:

TC> I guess I should clarify what I''m doing.

TC> Essentially I''d like to have the / and swap on the first 60GB of
TC> the disk.  Then use the remaining 100GB as a zfs partition to
TC> setup zones on.  Obviously the snapshots are extremely useful in such a
setup :)

TC> Does my plan sound feasible from both a usability and performance
standpoint?


Well, I''m doing very similar config in a production (to complete other
responses about their laptop usage).

It works great, no problems and usability... well, that''s probably
most important reason in such configs.


-- 
Best regards,
 Robert                            mailto:rmilkowski at task.gda.pl
                                       http://milek.blogspot.com

I''m thinking that if that is the case I''ll just be
dd''ing to a new disk and continuing on with it.  Obviously this is not
the preferred solution, but unless they''re willing to let me send my
own hardware, I don''t have much of a choice.
 
 
This message posted from opensolaris.org

Hello Al,

Thursday, January 25, 2007, 6:52:58 PM, you wrote:

AH> On Thu, 25 Jan 2007, Darren J Moffat wrote:
>> Tim Cook wrote:
>> > I guess I should clarify what I''m doing.
>> >
>> > Essentially I''d like to have the / and swap on the first
60GB of the
>> > disk.  Then use the remaining 100GB as a zfs partition to setup
zones
>> > on.  Obviously the snapshots are extremely useful in such a setup
:)
>> >
>> > Does my plan sound feasible from both a usability and performance
standpoint?
AH> For a laptop scenario, where you don''t have any option to add a
2nd drive,
AH> I agree with Darren and others.  But in any other usage scenario, I
don''t
AH> think its a good idea.  Perhaps you could post some details about your
AH> hardware environment and why its limited to a single disk drive?

AH> You understand that if ZFS detects any data corruption and you
don''t have
AH> any designed-in data redundancy, ZFS will make the pool unavailable and
AH> you are SOL.

It''s not true. Metadata will still be in 2 and 3 copies (2 for fs
metadata and 3 for pool metadata) spread on disks with checksums. So
when he will enocunter any problems to only part of a disk he should
be ok when it comes to pool/fs consistency and only user data will be
affected. However he should be able to determine which files are
affected (with zpool status) and just delete them. Few days ago a fix
was integrated so zpool status should show actual file names instead
of just inode.


-- 
Best regards,
 Robert                            mailto:rmilkowski at task.gda.pl
                                       http://milek.blogspot.com

Tim Cook wrote:
> I guess I should clarify what I''m doing.
>
> Essentially I''d like to have the / and swap on the first 60GB of
the disk.  Then use the remaining 100GB as a zfs partition to setup zones on. 
Obviously the snapshots are extremely useful in such a setup :)
>
> Does my plan sound feasible from both a usability and performance
standpoint?
>   
Yes, but depending on what you intend to do with zones, you
might get bit by this:

6356600 - zfs datasets can''t be used to provide space for zone roots

However, you actually CAN use zfs datasets for zone roots.
Works fine.  The reason for the bug is that patching and upgrade
don''t work yet on zone roots that are in zfs datasets, so we
don''t
recommend them for production use.  However, since it sounds
like you''re just setting this up for testing, this shouldn''t
be a problem
for you.

- Lori Alt

On Thu, 25 Jan 2007, Robert Milkowski wrote:
> Hello Al,
>
> Thursday, January 25, 2007, 6:52:58 PM, you wrote:
>
> AH> On Thu, 25 Jan 2007, Darren J Moffat wrote:
>
> >> Tim Cook wrote:
> >> > I guess I should clarify what I''m doing.
> >> >
> >> > Essentially I''d like to have the / and swap on the
first 60GB of the
> >> > disk.  Then use the remaining 100GB as a zfs partition to
setup zones
> >> > on.  Obviously the snapshots are extremely useful in such a
setup :)
> >> >
> >> > Does my plan sound feasible from both a usability and
performance standpoint?
>
> AH> For a laptop scenario, where you don''t have any option to
add a 2nd drive,
> AH> I agree with Darren and others.  But in any other usage scenario, I
don''t
> AH> think its a good idea.  Perhaps you could post some details about
your
> AH> hardware environment and why its limited to a single disk drive?
>
> AH> You understand that if ZFS detects any data corruption and you
don''t have
> AH> any designed-in data redundancy, ZFS will make the pool unavailable
and
> AH> you are SOL.
>
> It''s not true. Metadata will still be in 2 and 3 copies (2 for fs
> metadata and 3 for pool metadata) spread on disks with checksums. So
> when he will enocunter any problems to only part of a disk he should
> be ok when it comes to pool/fs consistency and only user data will be
> affected. However he should be able to determine which files are
> affected (with zpool status) and just delete them. Few days ago a fix
> was integrated so zpool status should show actual file names instead
> of just inode.
Thanks for the correction Robert.

Regards,

Al Hopper  Logical Approach Inc, Plano, TX.  al at logical-approach.com
           Voice: 972.379.2133 Fax: 972.379.2134  Timezone: US CDT
OpenSolaris.Org Community Advisory Board (CAB) Member - Apr 2005
             OpenSolaris Governing Board (OGB) Member - Feb 2006

does liveupgrade work fine if the zones are on a UFS partition?
 
 
This message posted from opensolaris.org

Tim Cook wrote:
> does liveupgrade work fine if the zones are on a UFS partition?
>  
>  
>   
Starting with build 53 (or was it 54?), yes.  Liveupgrade of systems
with zones is now supported. 

Lori

Thank you lori!  Guess I''ll take that route :)  The final solution is
actually a production environment.  For now I was going to stage, but it looks
like it''s not possible at this point in time.
 
 
This message posted from opensolaris.org

On 25-Jan-07, at 3:56 PM, Robert Milkowski wrote:
> Hello Tim,
>
> Thursday, January 25, 2007, 4:44:34 PM, you wrote:
>
> TC> I guess I should clarify what I''m doing.
>
> TC> Essentially I''d like to have the / and swap on the first
60GB of
> TC> the disk.  Then use the remaining 100GB as a zfs partition to
> TC> setup zones on.  Obviously the snapshots are extremely useful  
> in such a setup :)
>
> TC> Does my plan sound feasible from both a usability and  
> performance standpoint?
>
>
> Well, I''m doing very similar config in a production (to complete
other
> responses about their laptop usage).
Works fine for me on a production system too (root + swap slices SVM,  
big slice ZFS) but mirrored to 2nd disk with same layout.

--Toby
>
> It works great, no problems and usability... well, that''s probably
> most important reason in such configs.
>
>
> -- 
> Best regards,
>  Robert                            mailto:rmilkowski at task.gda.pl
>                                        http://milek.blogspot.com
>
> _______________________________________________
> zfs-discuss mailing list
> zfs-discuss at opensolaris.org
> http://mail.opensolaris.org/mailman/listinfo/zfs-discuss

Hello Toby,

Thursday, January 25, 2007, 10:38:16 PM, you wrote:

TT> On 25-Jan-07, at 3:56 PM, Robert Milkowski wrote:
>> Hello Tim,
>>
>> Thursday, January 25, 2007, 4:44:34 PM, you wrote:
>>
>> TC> I guess I should clarify what I''m doing.
>>
>> TC> Essentially I''d like to have the / and swap on the
first 60GB of
>> TC> the disk.  Then use the remaining 100GB as a zfs partition to
>> TC> setup zones on.  Obviously the snapshots are extremely useful  
>> in such a setup :)
>>
>> TC> Does my plan sound feasible from both a usability and  
>> performance standpoint?
>>
>>
>> Well, I''m doing very similar config in a production (to
complete other
>> responses about their laptop usage).
TT> Works fine for me on a production system too (root + swap slices SVM,
TT> big slice ZFS) but mirrored to 2nd disk with same layout.

TT> --Toby
>>
>> It works great, no problems and usability... well, that''s
probably
>> most important reason in such configs.
Of course mirrors :)

-- 
Best regards,
 Robert                            mailto:rmilkowski at task.gda.pl
                                       http://milek.blogspot.com

Hi,
> When you do the initial install, how do you do the slicing?
> 
> Just create like:
> / 10G
> swap 2G
> /altroot 10G
> /zfs restofdisk
yes.
> Or do you just create the first three slices and leave the rest of the
> disk untouched?  I understand the concept at this point, just trying to
> explain to a third party exactly what they need to do to prep the system
> disk for me :)
No. You need to be able to tell ZFS what to use. Hence, if your pool is
created at the slice level, you need to create a slice for it.

So the above is the way to go.

And yes, you only should do this on laptos and other machines where you only
have 1 disk or are otherwise very disk-limited :).

Best regards,
   Constantin

-- 
Constantin Gonzalez                            Sun Microsystems GmbH, Germany
Platform Technology Group, Client Solutions                http://www.sun.de/
Tel.: +49 89/4 60 08-25 91                   http://blogs.sun.com/constantin/

> Take note though, that giving zfs the entire disk gives a possible
> performance win, as zfs will only enable the write cache for the disk
> if it is given the entire disk.
really? 
why this?
is this tuneable somehow/somewhere? can i enabyle writecache if only using a
dedicated partition ?
 
 
This message posted from opensolaris.org

>> Take note though, that giving zfs the entire disk gives a possible
>> performance win, as zfs will only enable the write cache for the disk
>> if it is given the entire disk.
>
>really? 
>why this?
In the old days, Sun never enabled the write cache on devices because
of reliability issues.  (Sun SCSI disks were shipped with caches disabled;
but the OS never bothered to change the caching behaviour; the behaviour
on SCSI drives is persistent.

On ata drives, the drive cache was specifically disabled (the behaviour
is not persistent and drives default to write cache on)

This behaviour was changed under competetive pressure for sata disks;
they now default to write cache on; set using the
"sata:sata_write_cache"
variable

The change came about with ZFS and the addition of a mechanism to
flush the write cache (ZFS needs this to guarantee transactional
safety)(
>is this tuneable somehow/somewhere? can i enabyle writecache if only using a
dedicated partition ?
If does put the additional data at some what of a risk; not really
for swap but perhaps not nice for UFS.

Casper

On Sun, Jan 28, 2007 at 01:53:04PM +0100, Casper.Dik at Sun.COM
wrote:
> 
> >is this tuneable somehow/somewhere? can i enabyle writecache if only
using a dedicated partition ?
> 
> If does put the additional data at some what of a risk; not really
> for swap but perhaps not nice for UFS.
How about two partitions used in two different ZPOOLs?  once ZFS boot comes
along, I''m sure RAIDZ(2) won''t be supported as a boot device? 
If that''s the
case, I wouldn''t mind splitting the disks into a mirrored OS portion
and a
RAIDZ data portion (think system with 3 or 4 disks).

-brian
-- 
"The reason I don''t use Gnome: every single other window manager I
know of is
very powerfully extensible, where you can switch actions to different mouse
buttons. Guess which one is not, because it might confuse the poor users?
Here''s a hint: it''s not the small and fast one."       
--Linus