Text 2816, 59 rader
Skriven 2011-02-23 21:05:28 av Jonathan de Boyne Pollard
Ärende: Re: Ubuntu installation modifies the partition table rendering IBM LVM 
===============================================================================
Gecko/20101207 Thunderbird/3.1.7
UTC)
comp.os.linux.setup:7797 comp.os.os2.setup.storage:381
From: Jonathan de Boyne Pollard <J.deBoynePollard-newsgroups@NTLWorld.COM>

 I guess my real point was that this problem isn't specific to LVM, 
> although it's possibly specific to OS/2 in general.
>
Well this particular problem *is* specific to LVM.

>> It's what IBM's LVM is doing that isn't fine. It's making unwarranted 
>> assumptions about where it can just grab an unallocated disc block 
>> without explicitly allocating it. As I noted, other disc management 
>> systems *do not do things this way*. Microsoft's Logical Disk Manager 
>> uses a special partition, whose type is "LDM Metadata" (EFI partition 
>> type 5808c8aa-7e8f-42e0-85d2-e1e90434cfb3), and puts the metadata 
>> there. The Linux LVM (we are told) stores the metadata within each 
>> partition. There's a reason that the rest of the world doesn't do 
>> things the IBM LVM way, and hasn't done for at least a decade now 
>> (There is Windows 2000 Server doco that talks about the pitfalls of 
>> expecting unallocated areas of the partition table to be usable.), 
>> and this experience *is* that reason. It's a flawed and selfish 
>> design and doesn't work.
>>
> Begs the question of course... what, if anything, can be done about it?
>
> I suppose some ambitious developer could rewrite LVM, although when 
> IBM open-sourced it for the Linux people to use, they only seem to 
> have released half of it: LVM.DLL was open-sourced, but not (AFAIK) 
> OS2DASD.DMD and OS2LVM.DMD, which AIUI do fairly critical parts of the 
> work...
>

They do for OS/2.  But they aren't necessary for another operating 
system, since the device driver model of, say, Linux is quite different 
to that of OS/2.  What one needs in such circumstances are good and 
thorough specifications for the on-disc data structures, which I presume 
that there were.  From that one can write a driver.

It's worth noting, too, that drivers may are not strictly necessary for 
other operating systems that already have LVM mechanisms.  The Linux LVM 
model, we are told by the not-very-reliable-when-it-comes-to-computers 
Wikipedia, is for the kernel to only know how to stitch multiple 
discontiguous disc slices together to make a single volume, that a 
filesystem driver then sits on top of.  All of the working out exactly 
*what* slices to stitch together is the province of applications, not 
the kernel.  Follow through on that model, and support for IBM LVM 
becomes a simple case of parsing the on-disc LVM metadata that describes 
where the disc slices are, and issuing the relevant system calls to tell 
the Linux kernel what to stitch to what.  Of course it's not *quite* 
that simple for Linux in reality, but the principle is a general one.

For IBM OS/2, yes one does need replacement DMDs, because it's the DMDs 
that go looking at partition tables and decide what disc slices are what 
on a disc.


--- Internet Rex 2.31
 * Origin: virginmedia.com (1:261/20.999)