[Way OT] dieresis

On Tue, 09 Mar 2010 20:48:34 -0800, JosephKK wrote:

I use parted as needed. Not so skillful with dd yet. "cp -R *" and the
like serve me pretty well.

cp -R * won't transfer the boot sector, partition table, etc.

My typical usage is to partition the new disk the way i want and copy the

Just use "dd if=/dev/hda of=/dev/hdc" with the target drive on the second
IDE. Then go to bed. It'll be finished in the morning :-)

And the sizes of the new partitions mirror the old ones. I don't seem to

Don't do it with mounted drives. Boot with a startup floppy or CD with a
copy of dd and parted on it.

Don't worry about BIOS settings. Linux
doesn't use them for I/O. I've actually run a 40 gig HDD on a machine with
an 8 gig BIOS limit, (with a 1024 cylinder boot partition to fool it).

You'll find all your original partitions recreated on your new drive, with
empty space left if it's bigger. It'll boot, too.

Then use parted to stretch the created partition to the size of the new
(bigger) HDD. Go for breakfast whilst that's running.

Thanks for the offer of the monitors. I have quite a few spare already.

And the sizes of the new partitions mirror the old ones. I don't seem to
have come across a version of parted that will just do as i ask, instead
of second guessing me.

Fer instance; stretch the extended partition to the end of the disk, move
the internal partitions to the end, move the beginning of the extended
partition, modify existing primary partitions. Most bitch about the first
change and only one so far did the second. All versions i have tried
refused to do the third task. Rather non-*nix like i say.

On Thu, 11 Mar 2010 21:35:45 -0800, JosephKK wrote:

And the sizes of the new partitions mirror the old ones. I don't seem to
have come across a version of parted that will just do as i ask, instead
of second guessing me.

Fer instance; stretch the extended partition to the end of the disk, move
the internal partitions to the end, move the beginning of the extended
partition, modify existing primary partitions. Most bitch about the first
change and only one so far did the second. All versions i have tried
refused to do the third task. Rather non-*nix like i say.

I've never used extended partitions. Four primary to a drive is enough for
me. In fact three primary partitions are all I have, now: boot - swap -
root, with the boot partition (/dev/hda1) mounted on /boot in /etc/fstab
at startup.

I've never had any trouble with parted not doing what I wanted. I'll dig
the manual out and refresh my memory on what it will/won't do. My version
is old (1.6.3)

frown> I don't run Linux so keep that in mind...

Can't you (or, *don't* you?) build slices *within* a "partition"? E.g., I
can set up an x86 PC to use one of the (4) partitions for DOS, another for
some variant of Windows a third for NetBSD 3.1 and the fourth for NetBSD
5.0.

DOS and Windows are happy -- since they deal with "partitions".

Within each of the NetBSD partitions, I create many (4, 8, 10?) "slices"
(think: partition : disk :: slice : partition) and each of those is
(potentially) a filesystem mounted someplace in the (NetBSD) hierarchy.

E.g., /, /var, /usr, /usr/local, /usr/pkg, /home ...

On Fri, 12 Mar 2010 14:58:03 -0700, D Yuniskis wrote:

frown> I don't run Linux so keep that in mind...

Can't you (or, *don't* you?) build slices *within* a "partition"? E.g., I
can set up an x86 PC to use one of the (4) partitions for DOS, another for
some variant of Windows a third for NetBSD 3.1 and the fourth for NetBSD
5.0.

DOS and Windows are happy -- since they deal with "partitions".

I don't run DOS or Windows. Strictly Microsoft-free zone here. Just Linux.

Within each of the NetBSD partitions, I create many (4, 8, 10?) "slices"
(think: partition : disk :: slice : partition) and each of those is
(potentially) a filesystem mounted someplace in the (NetBSD) hierarchy.

E.g., /, /var, /usr, /usr/local, /usr/pkg, /home ...

I don't know anything about NetBSD filesystems. Linux fdisk will make
NetBSD partitions, but I don't know whether there's an mkfs for NetBSD.

I used to have /usr and /home on a separate physical drive, once, when all
I had were smallish drives. I can't see any point in having lots of
separate partitions otherwise. My boot partition is a relic of the days of
BIOS that would only boot from the first 1024 cylinders.

Fred Abse wrote:
On Fri, 12 Mar 2010 14:58:03 -0700, D Yuniskis wrote:

frown> I don't run Linux so keep that in mind...

Can't you (or, *don't* you?) build slices *within* a "partition"? E.g.,
I can set up an x86 PC to use one of the (4) partitions for DOS,
another for some variant of Windows a third for NetBSD 3.1 and the
fourth for NetBSD 5.0.

DOS and Windows are happy -- since they deal with "partitions".

I don't run DOS or Windows. Strictly Microsoft-free zone here. Just
Linux.

Doesn't matter. I could make them all {Net,Open,Free}BSD "partitions"...

Within each of the NetBSD partitions, I create many (4, 8, 10?)
"slices" (think: partition : disk :: slice : partition) and each of
those is (potentially) a filesystem mounted someplace in the (NetBSD)
hierarchy.

E.g., /, /var, /usr, /usr/local, /usr/pkg, /home ...

I don't know anything about NetBSD filesystems. Linux fdisk will make
NetBSD partitions, but I don't know whether there's an mkfs for NetBSD.

I used to have /usr and /home on a separate physical drive, once, when
all I had were smallish drives. I can't see any point in having lots of
separate partitions otherwise. My boot partition is a relic of the days
of BIOS that would only boot from the first 1024 cylinders.

Having partitions is a double-edged sword. Put everything in / and if
something corrupts that file system, you're hosed. OTOH, keep / small
and effectively R/O, then you can be reasonably sure that you can mount
it (single user) and, from there, troubleshoot problems you might have
with other "not yet mounted" file systems.

You can also treat them as different file systems (with different
characteristics).

(there's a long list of the "cans" as well as the "downsides")

When I first started sysadm'ing UN*X boxes (before the free eunices came
into being), I found partitions to be a PITA. Over time, my attitude
has flipped 180 on this issue...