Re: [Bulk] Re: Boot configuration. Was Re: [opensuse] Internet monitoring tool
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 El 2010-03-02 a las 17:20 -0500, Anton Aylward escribió: You forgot to email to the list.
Carlos E. R. said the following on 03/02/2010 04:47 PM:
No, grub does not reorder. Number five is number five regardless the order it has.
We're talking at cross purposes. Let me try again.
I ran 'fdisk' to get a listing of the partitions. Fdisk orders by "sda*" and not by the order of the partitions
I recorder the output of 'fdisk' to match the ascending order of the cylinders.
If the order of the cylinders is not that of the partitions ... then something is wrong.
Nothing is wrong. They are different, that's all. Don't reorder, simply print whatever fdisk prints, because it is correct. That is the order of the partitions, not the tracks.
Go back and count.
I don't need to count. Partition 6, /dev/sda6, is always the second logical partition, by definition.
No, sda5 is the first, always. Cylinder order is irrelevant.
Are you talking about order of creation? No, that can't be right.
I'm talking that the partition that gets the number 5 by fdisk is the first logical partition, regardless of whether it is placed the first in the track layout or the last. Look: don't look at the tracks numbers, forget it. Fdisk says "5", linux says "/dev/sda5". That's all that matters. The software knows perfectly well haw to access partition numbered "5".
This isn't making any sense. If the order of the cinders is irrelevant what is the underlying logic.
The partition numbers depends on where in a table is each partition listed. There are four entries, the first one is numbered "one" - even if it is physically the last partition in the disk. The table is boss, pysical layout is arbitrary. Kind of.
Are you saying that the whole thing gets re-written on every edit?
No.
Don't bother about the order the partitions are defined in the disk, the important thing is the numbers assigned to them. And the sofware will not get confused by that. Only us ;-)
Well that last statement is certainly true!
:-)
(hd0,5) is /dev/sda6. Correct.
Now since grub, sorry, GRUB, counts from zero, why is this "5" and not "4"?
Ah.
(hd0,5) sda6 (hd0,4) sda5 (hd0,3) sda4 (hd0,2) sda3 (hd0,1) sda2 (hd0,0) sda1
That's programmesse >:-)
Double DUH
Must be a C programmer.
Or assembler. - -- Saludos Carlos E. R. -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.9 (GNU/Linux) iEYEARECAAYFAkuNsbQACgkQtTMYHG2NR9WHiQCgiHdOAcX8btDfZY6j5FjB5QOd eWoAn2nmOnjtFbSAny9xwPHg81WPBcGY =r7Bp -----END PGP SIGNATURE-----
Carlos E. R. said the following on 03/02/2010 07:47 PM:
This isn't making any sense. If the order of the cinders is irrelevant what is the underlying logic.
The partition numbers depends on where in a table is each partition listed. There are four entries, the first one is numbered "one" - even if it is physically the last partition in the disk. The table is boss, pysical layout is arbitrary. Kind of.
Are you saying that the whole thing gets re-written on every edit?
No.
But surely it has to. When I shrank the second partition and put a new one in there ... Are you saying the table entry existed anyway? Its not about a table, its a bout 'slots'? Does it have some way of telling zero from a blank entry? (larry wall apropos quote goes here) Well a 'slots' approach begs the question about the number of slots for the partition in the extended partition. Only one extended partition, you say ...
And the sofware will not get confused by that. Only us ;-)
Well that last statement is certainly true!
:-) -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 On Tuesday, 2010-03-02 at 20:22 -0500, Anton Aylward wrote:
Carlos E. R. said the following on 03/02/2010 07:47 PM:
This isn't making any sense. If the order of the cinders is irrelevant what is the underlying logic.
The partition numbers depends on where in a table is each partition listed. There are four entries, the first one is numbered "one" - even if it is physically the last partition in the disk. The table is boss, pysical layout is arbitrary. Kind of.
Are you saying that the whole thing gets re-written on every edit?
No.
But surely it has to. When I shrank the second partition and put a new one in there ...
Are you saying the table entry existed anyway? Its not about a table, its a bout 'slots'?
In your case the entry existed. And it is a table. :-) I explain: entries 1, 2, 3 and 4 are fixed, they exist always. They are "primaries". Entries from 5 onward (logical partitions) are a linked list: you add entries to the end of the list, but new entries can, nevertheless, point to an area in the disk at the very start. Or to the middle. The position of the disk space assigned to a partition is not related to the partition number. So, when you "add" partition number 3, you are simply filling up entry number 3, and writing in there where the space assigned to that partition is. By the way, reordering the entries means, in fact, erasing all the entries and writing them all again, with the modifications. A wrong change and you can no longer access your data... so we don't do that.
Does it have some way of telling zero from a blank entry? (larry wall apropos quote goes here)
Obviously there is a method to signal an unused entry. I don't know the table format, I haven't needed to. But if you are curious, it is here: http://en.wikipedia.org/wiki/Master_boot_record#MBRs_and_disk_partitioning
Well a 'slots' approach begs the question about the number of slots for the partition in the extended partition. Only one extended partition, you say ...
Quick answer: unlimited. Detailed answer: logical partition "table" is not a table, but a linked list, and thus, unlimited. However, the operating system may set a limit. In linux, for ide disks (hda, hdb, hdc...) there was a limit of 63, IIRC. For scssi (sda, sdb, sdc...) the limit was 15, on versions prior to oS 11.2, and unlimited (I think) from 11.2 onwards. - -- Cheers, Carlos E. R. -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.9 (GNU/Linux) iEYEARECAAYFAkuOzIsACgkQtTMYHG2NR9UPiACghF89GyBQxC/70C4mqepbEWyb FAUAn0RNMtzB/+Fl1WJCTYCwqwTBsALt =+0oF -----END PGP SIGNATURE----- -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
participants (2)
-
Anton Aylward -
Carlos E. R.