Re: Partition limit on Solaris disk drives

According to Ian Collins <ian-news@xxxxxxxxxxx>:
DoN. Nichols wrote:
According to Ian Collins <ian-news@xxxxxxxxxxx>:
DoN. Nichols wrote:

[ ... ]

On the boot disk? I thought that ZFS will only work for
filesystems other than the root filesystem -- that you have to be enough
through the boot sequence before it will create and mount the
filesystems.

As Rich pointed out, you can use ZFS on a slice, the layout he described
is exactly what I use on my laptop. I can't think of any reason why one
wouldn't use ZFS for /export.

Well -- aside from not *having* a /export partition. :-) The
largest partition as I have it set up is /dumps -- for use by amanda.

I'm already experimenting with zfs and a cluster of 18 GB SCSI
drives, and once I have enough FC disks of sufficient size, then yes,
ZFS will be used for that. But since it seems to be at its most
efficient, I don't really want to make only part of the primary (boot)
disk ZFS while the rest is plain filesystems for boot purposes.

Why not? It works well and you get the ZFS end to end checksumming.
You can create as many filesystems in your pool with whatever properties
that best suit their contents.

O.K. I though that you paid a speed penalty for using a single
slice or a file instead of the whole disk.

From the man page for zpool:


======================================================================
disk

A block device, typically located under "/dev/dsk". ZFS
can use individual slices or partitions, though the
recommended mode of operation is to use whole disks. A
disk can be specified by a full path, or it can be a
shorthand name (the relative portion of the path under
"/dev/dsk"). A whole disk can be specified by omitting
the slice or partition designation. For example,
"c0t0d0" is equivalent to "/dev/dsk/c0t0d0s2". When
given a whole disk, ZFS automatically labels the disk,
if necessary.
======================================================================

They don't say there *why* the recommended mode is to use whole
disks, and I see that I was wrong about files within filesystems being
possible devices for zpool to use.

It does take a while to get into the ZFS mind set.

Of course.

But I want to benefit from either mirroring or ideally RAIDZ
so that wants identical partitions or disks. (That is why I am using a
collection of older 18 GB drives in an old Multipack, to assure that I
can feed all of the devices as the same size so the RAIDZ works
properly.

Thanks,
DoN.

--
Email: <dnichols@xxxxxxxxxxx> | Voice (all times): (703) 938-4564
(too) near Washington D.C. | http://www.d-and-d.com/dnichols/DoN.html
--- Black Holes are where God is dividing by zero ---
.

Relevant Pages

  • Re: [SLE] 9.0 Device checling on ext3 fs
    ... run the badbloc program on all partitions once. ... With something as critical as disk ... > filesystems, I need an app that has been tested and just works, preferably ... Usually they come as a file to create an MsDos boot ...
    (SuSE)
  • Re: Filesystem snapshots dog slow
    ... small filesystems, that are not busy. ... your disk is hammered making copies of all the cylinder ... groups, skipping those that are 'busy', and coming back to them later. ... This doesn't mean ZFS is ...
    (freebsd-hackers)
  • Re: Problem on boot
    ... > disk caches and synchronizes the disks such that there aren't any ... Having said that, both ext3 and reiserfs are journalled filesystems, so ... filesystems or ones that are recreated on each boot. ... James Wilkinson ...
    (Fedora)
  • Re: Partition limit on Solaris disk drives
    ... the seven plus the whole disk which Solaris allows. ... Use ZFS filesystems. ... drives, and once I have enough FC disks of sufficient size, then yes, ...
    (comp.sys.sun.hardware)
  • Re: Partition limit on Solaris disk drives
    ... the seven plus the whole disk which Solaris allows. ... Use ZFS filesystems. ... drives, and once I have enough FC disks of sufficient size, then yes, ...
    (comp.sys.sun.hardware)