Re: Status of Debian on Sun Blade 150 (parted and the like)

2005-04-29 Thread Frans Pop
On Friday 29 April 2005 15:47, Wiktor Wandachowicz wrote:
 So, after all it looks like the problem was on my side. I'm terribly
 sorry for misinforming you, and I take back my words. Looks like Frans
 was right when he wrote about user error. I bow my head before him. He
 is the one who really knows better how to handle Debian on sparc... But
 I'm learning :-) 

:-)   /me is only a newbie regarding Sparc himself ;-)

 Probably putting swap partition in a different location could save me a
 bit of trouble. But on the side note, it was Solaris installer that put
 swap partition on cylinders 0-258 of the hard drive. So it looked to me
 that it knew what it was doing proposing such layout upon install.

It probably does for Solaris.
I think that Solaris may format/use the swap partition in a different way 
than linux (and thus Debian) does.

So, the problem is that the installer blindly re-uses (and also formats 
by default) a swap partition that starts in sector 0.
There already is some code and dialogs in silo-installer that hooks into 
partman to warn about this, but that code appears to be unfinished and is 
currently not used.

Cheers,
FJP


pgpNkeMTxbmkT.pgp
Description: PGP signature


Re: Status of Debian on Sun Blade 150 (parted and the like)

2005-04-29 Thread Wiktor Wandachowicz
Frans Pop wrote:

Wiktor Wandachowicz wrote:

Probably putting swap partition in a different location could save me a
bit of trouble. But on the side note, it was Solaris installer that put
swap partition on cylinders 0-258 of the hard drive. So it looked to me
that it knew what it was doing proposing such layout upon install.


It probably does for Solaris.
I think that Solaris may format/use the swap partition in a different way 
than linux (and thus Debian) does.

When we'll be able to see the sources of OpenSolaris, it may as well become
quite obvious. But I'd like to repeat: on running Debian system I've put
a script which recreates swap every time Linux boots (btw. Solaris
initializes swap itself, without any action needed).

The sript I've made is in /etc/init.d/regenswap.sh, and is symlinked to
/etc/rcS.d/S09regenswap.sh. This script essentially calls mkswap for every
swap partition listed in /etc/fstab, BEFORE any swapon is called.
This way the swap partition is shared between Linux and Solaris and there
is no need to create an additional partition for linux-style swap only.

So, the problem is that the installer blindly re-uses (and also formats 
by default) a swap partition that starts in sector 0.
There already is some code and dialogs in silo-installer that hooks into 
partman to warn about this, but that code appears to be unfinished and is 
currently not used.

I don't know if it could stop me, though. I'm quite determined sometimes :-)

Friendly,
Wiktor Wandachowicz


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]