On 2007/09/20 21:00 (GMT+0200) Andreas Jaeger apparently typed:
Felix Miata wr0te:
On 2007/09/20 00:44 (GMT-0600) bugzilla_noreply@novell.com apparently typed:
--- Comment #25 from Andreas Jaeger
2007-09-20 00:44:01 MST --- See the release notes
They don't say what one should find in /dev as a result of running activate_dm_linear or what if anything people running >15 on SATA can do or expect.
Ignore activate_dm_linear, it's really not working.
Btw. you even get a popup which advises to use "hwprobe=-modules.pata".
I haven't seen that yet, but I haven't been past the first YaST screen in over a month.
Tentative release notes for 10.3 seem to be nicely hidden, not yet on http://en.opensuse.org/Release_Notes. It would be nice to see http://www.suse.com/relnotes/i386/openSUSE/10.2/RELEASE-NOTES.en on http://en.opensuse.org/Factory and/or http://en.opensuse.org/Development_Version
(and bug # 305095)-
I've been following that and every other related bug (keyword >15). Most of what's in them seems pretty cryptic. I would expect running activate_dm_linear /dev/sda at the start of an RC1 install should produce some evidence that anything happened in /dev, but I don't see it.
As I've said: It misses some files in the inst.sys and therefore just does not work in RC1.
If that's referring to https://bugzilla.novell.com/show_bug.cgi?id=326692 I'd like some notification when it hits the factory mirrors instead of waiting over a week for RC2.
and use "hwprobe=-modules.pata" to not use libata at all.
I don't see how that could be of any help on SATA systems with >15 partitions. I already use it on PATA.
A SATA system with > 15 partitions was never ever supported!
I well understand that. However https://bugzilla.novell.com/show_bug.cgi?id=218122 summary says "Support more than 15 partitions on libata", not "Support more than 15 partitions on PATA on libata". Unless I grossly misunderstand the nature of libata, I expect libata to mean libata HD support regardless whether used for PATA or SATA. If I'm not wrong in my understanding, and yet >15 on SATA is unlikely to be supported now or within next year or so, then the summary of that bug seems to be wrong, and should be changed to remove any implication of applicability to SATA either now or in the future.
If you want that, then use LVM.
That may be fine for most people, as that's apparently what the kernel developers expect, as well as the only offering for Fedora users. It's not so good for people with a tried and true multiboot creation, maintenance, backup and restore strategy based upon cloning/copying partitions and a minimal number of physical hard disks per system. LVM is no solution to many of us with true multiboot of multiple versions of Linux and other operating systems (more than ~3 total). LVM is not OS agnostic. On systems that use it, systems that don't understand it have to work around it, which to my knowledge as a user of >15 on multiple systems for many years is simply not practical. IOW, if >15 is never to be supported on SATA, those of us committed to >15 are committed to having to buy converter gadgets to use SATA drives as PATA drives as the old PATA drives need replacement[1], and to keeping legacy systems operational well beyond expected lifetimes, since new motherboards omit PATA controllers, and PATA add-in cards are treated as SCSI cards not under boot order control of the PC system BIOS. It's really absurd that as HD sizes continue to escalate that users should be forced to using ever stupidly larger partitions due to the kernels arbitrary limitation of 14 replacing the traditional 62. A limit of 14 on a 500GiB drive means either average partition size over 10 times the size of my current average partition size, or wasting 80% or more of the disk to keep partitions to a manageable size for backup and pruning purposes. [1] Seagate, world's largest HD supplier, has already announced imminent cessation of PATA drive production.
the activate_dm_linear solution would only help with updating existing PATA systems.
Does this mean it could not be applied to SATA, or merely that was not the intent of creating a solution for upgrading systems with >15 on PATA?
The workaround from Hannes needs some more work we will do for RC2 - but even then it will not be nicely integrated and only something for a real hacker.
From what I can tell in the relevant bugs, activate_dm_linear seems simple enough for an unreal hacker to manage. I've been ready for quite some time if only I could find functional guidance on what to do or evidence that anything resulted from what I did.
This needs more work for 11.0.
Obviously. :-p I feel like the progress so far is behind a locked door to which the public key is hidden by secret code, or doesn't exist.
The general advise really is to use hwprobe=-modules.pata - this works just fine.
On Mandriva Cooker 2008 it's conceptually simpler, even though it requires an extra step, because it's easier to remember. On the kernel line one appends the six letters "noauto", after which the installer allows explicit selection of the driver to be used for PATA.
On SATA?
What is FATE?
FATE is our FeAture Tracking Tool
Is it, as it applies to OpenSUSE, Novell internal access only? If not, an URL to it's OpenSUSE-applicable home would be nice. On 2007/09/20 21:01 (GMT+0200) Andreas Jaeger apparently typed:
Carlos E. R. wrote:
The Thursday 2007-09-20 at 09:19 -0400, Felix Miata wrote:
They don't say what one should find in /dev as a result of running activate_dm_linear or what if anything people running >15 on SATA can do or expect.
I think they would appear under /dev/mapper/. I have used the device mapper thing previously (for encrypted filesystems in old twofishSL92 format), but I know pretty little about it, close to nothing. I believe it can be a very interesting piece of gadgetry, but I haven't seen any user documentation. A howto.
Hannes promised me to write a howto in the openSUSE wiki.
Excellent. Thanks in advance.
But with RC1 it just does not work because two binaries are missing in the inst.sys,
I'm anxiously awaiting a factory mirror sync that includes this. Lack of ability to test this has previously just missed beta/rc release cuts a few times already. It seems Hannes has been the only one in a position to even try out his work. -- "It yet remains a problem to be solved in human affairs, whether any free government can be permanent, where the public worship of God, and the support of religion, constitute no part of the policy or duty of the state in any assignable shape." Chief Justice Joseph Story Team OS/2 ** Reg. Linux User #211409 Felix Miata *** http://mrmazda.no-ip.com/ --------------------------------------------------------------------- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org For additional commands, e-mail: opensuse-factory+help@opensuse.org