Reactivation of a boot menu after installation of openSUSE 12.1

This happened already by the openSUSE installation system and the tool “YaST boot loader setup”.

I updated it also in the meantime.

(hd0)   /dev/disk/by-id/ata-SAMSUNG_SP2504C_S09QJ1UA109248
(hd1)   /dev/disk/by-id/ata-SAMSUNG_SP2504C_S09QJ1UA109249
(hd2)   /dev/disk/by-id/ata-SAMSUNG_HD103SJ_S246J90B901367

A bit more information:

Sonne:~ # findgrub -M
Find Grub Version 3.5.1 - Written for openSUSE Forums

--- DEVICE.MAP: sda is ata drive hd0
--- DEVICE.MAP: looking for /dev/sda in /boot/grub/device.map:
--- DEVICE.MAP: - /dev/sda
--- DEVICE.MAP: - /dev/disk/by-id/ata-SAMSUNG_SP2504C_S09QJ1UA109248 -> hd0
--- DEVICE.MAP: - /dev/disk/by-id/scsi-SATA_SAMSUNG_SP2504CS09QJ1UA109248
--- DEVICE.MAP: - /dev/disk/by-id/wwn-0x50000f0016109248
--- DEVICE.MAP: - /dev/disk/by-path/pci-0000:00:09.0-scsi-0:0:0:0
--- DEVICE.MAP: => sda - found in device.map - is now hd0
--- DEVICE.MAP:
--- DEVICE.MAP: sdb is ata drive hd1
--- DEVICE.MAP: looking for /dev/sdb in /boot/grub/device.map:
--- DEVICE.MAP: - /dev/sdb
--- DEVICE.MAP: - /dev/disk/by-id/ata-SAMSUNG_SP2504C_S09QJ1UA109249 -> hd1
--- DEVICE.MAP: - /dev/disk/by-id/scsi-SATA_SAMSUNG_SP2504CS09QJ1UA109249
--- DEVICE.MAP: - /dev/disk/by-id/wwn-0x50000f0016109249
--- DEVICE.MAP: - /dev/disk/by-path/pci-0000:00:09.0-scsi-2:0:0:0
--- DEVICE.MAP: => sdb - found in device.map - is now hd1
--- DEVICE.MAP:
--- DEVICE.MAP: sdc is ata drive hd2
--- DEVICE.MAP: looking for /dev/sdc in /boot/grub/device.map:
--- DEVICE.MAP: - /dev/sdc
--- DEVICE.MAP: - /dev/disk/by-id/ata-SAMSUNG_HD103SJ_S246J90B901367 -> hd2
--- DEVICE.MAP: - /dev/disk/by-id/scsi-SATA_SAMSUNG_HD103SJS246J90B901367
--- DEVICE.MAP: - /dev/disk/by-id/wwn-0x50024e920600ae32
--- DEVICE.MAP: - /dev/disk/by-path/pci-0000:00:09.0-scsi-3:0:0:0
--- DEVICE.MAP: => sdc - found in device.map - is now hd2
--- DEVICE.MAP:
********************************************************************************

By the way: Do you find the details from the bug report ‘Does “cfdisk” not agree to cylinder alignment from partitioning by openSUSE installation system?’ interesting?

Would it make sense to try out GRUB2 on the first sector of my new hard disk?

Can the versions 0.97-177.1.2 and 1.99-8.7.2 be installed in parallel for the software component “Grand Unified Boot Loader”?

This post should answer your question better than I would have done:
Ubuntu Forums - View Single Post - [all variants] Partition does not end on cylinder boundary

I found a mention of CIFS improvements in the new kernel 3.2:

A number of changes to the CIFS filesystem code for accessing Samba or Windows shares should significantly improve data throughput in certain situations (1, 2, 3).

Located here: Miscellaneous, Minor gems - The H Open Source: News and Features

Thank You,

Not if you can not resist to press F8 before booting. But what are you trying to perform “exactly”? What is the expected result?

What do you mean by in parallel? The first one refers to Legacy Grub (still used in openSUSE). The second one refers to Grub2, used today in most other distros. They are different programs. You can use them both, just like you can install several OSes. People who dual-boot Ubuntu and openSUSE have at least one instance of each. Grub2 should be installed in MBR. Installing it in a partition bootsector is possible but discouraged and usually works until the first brutal fsck.

I do not recommend to install them both under the same installation of openSUSE … unless you intend to join the fun and help me to debug findgrub.

James, if it’s still too early in Texas, you might try this command first:

echo "CIFS" | sed 's/C\(.\)\(.\)\(.\)/C\2d\1\3k/' | tr ":upper:]" ":lower:]"

lol!

So the output of this command is just: cfdisk
It is kind of early, I have had only one cup of coffee and it is way to early to figure out any CLI commands you could put together. I started getting a headache just looking at that command line. :stuck_out_tongue:

Thank You,

On 2011-12-04 15:26, jdmcdaniel3 wrote:
>
> I found a mention of CIFS improvements in the new kernel 3.2:
>
>> A number of changes to the CIFS filesystem code for accessing Samba or
>> Windows shares ‘should’
>> (http://thread.gmane.org/gmane.linux.kernel.cifs/4696) significantly
>> improve data throughput in certain situations (‘1’
>> (http://git.kernel.org/linus/e28bc5b1fdbd6e850488234d6072e6b66fc46146),
>> ‘2’
>> (http://git.kernel.org/linus/9ee305b70e09f5132c9723780ce10e69710b8bca),
>> ‘3’
>> (http://git.kernel.org/linus/32b9aaf1a53b3c8d435f86339b01b3968520cb0a)).
>
> Located here: ‘Miscellaneous, Minor gems - The H Open Source: News and
> Features’ (http://tinyurl.com/cclph6j)

I’m baffled.

This thread is about a problem with grub. How does CIFS relate to this? :-?


Cheers / Saludos,

Carlos E. R.
(from 11.4 x86_64 “Celadon” at Telcontar)

I’m sorry. - I do not see how this link belongs to my story about boot menus.

I would appreciate if booting from this BIOS menu will also work.

But what are you trying to perform “exactly”?

I would like to assure that the blinking white underscore is displayed by GRUB on my screen so far after the F8 selection. Would it make sense to put anything into the MBR of the disk “/dev/sdc” that I left empty?
Do I need to consider that any other boot loader gets started from this place accidentally while I expect that only the second partition should be bootable from this disk?

What do you mean by in parallel?

I assume if I would try out the package “grub2” (version 1.99-8.7.2) that it might overwrite the previous GRUB1 configuration. (I do not want that at the moment.) I imagine to put it eventually on a different partition (the first?) so that an other display can give me a better feedback.

Was my explanation not clear enough?
All right, you have 2 possibilities, not 3 - Tertium non datur!

  • Either you forget about your BIOS menu and enter the funny world of multi-booting. In this case, welcome on board! What you are doing is the contrary of multi-booting. This is the reason why boot managers don’t do what you’re expecting.
  • Or you prefer to switch hard disks BIOS order, which is possible after all in a multicultural world, but in this case, you have to make each operating system aware that it is installed on the first hard disk.

Then you have to ask yourself if BIOS passed control to a disk containing a valid bootloader. Otherwise that underscore will be the answer of the BIOS itself, not able to do anything else. However the BIOS should display a “no operating system found” error, and Grub has a bunch of error codes for most situations. What’s sure is that someone doesn’t know what to do.

Yes, it would make sense. But what would make even more sense would be to not switch boot order with 3 hard disks installed.
I suspect what happens is that you’re trying to boot from a disk with an empty MBR and no active partition.

You need to keep in mind that many cheap and bad designed mainboards don’t handle 1-2-3 boot order as easily as 1-2. Further, either the bootloder in MBR of what is currently the boot disk will get start - provide there is one - or the bootloader installed in the active partition if there is one. If neither the MBR is bootable nor a partition set active, the BIOS might pass control to the next hard disk - but which will be the second one if you switch 1-2-3 randomly - assuming you did set a second boot disk in the BIOS setup, but that in turn might get messed up when you press F8.

Don’t install Grub2 in a partition bootsector! Read my posts on that topic or the Grub2 documentation.

Thanks for your clarification.

But what would make even more sense would be to not switch boot order with 3 hard disks installed.

I do not need to fiddle with BIOS boot priorities for these disks again as long as the menus from the boot managers provide appropriate access to the required operating systems and tools.

I suspect what happens is that you’re trying to boot from a disk with an empty MBR and no active partition.

I tried this use case with the second disk intentionally. It seems that the BIOS chooses to continue booting from the first disk.

You need to keep in mind that many cheap and bad designed mainboards don’t handle 1-2-3 boot order as easily as 1-2.

Do you know any more bug reports that are similar to my issue?

Yes. You should set the boot order once, depending on the OSes you want to boot, and then let the bootmanager do its job.

It depends on the hard disks you have set in BIOS setup for booting - you have some influence on that - and it depends on how many hard disks or other devices the BIOS let you include in the list (usually 3 or 4).

I fail to understand what is your problem exactly. If you boot a hard disk which doesn’t contain a valid bootloader and you see a blinking underscore, it’s not a bug, it’s normal. So what would be the bug here? If you have 3 hard disks and a CD ROM drive, and the mainboard will only boot 2 hard disks and the CD ROM drive, it’s a BIOS limitation. Same for beeing able to boot or not from USB Flash or external drives. If you add a third hard disk in the third SATA port, set the BIOS order properly or not a all - so that it would just follow the order of the SATA ports, and the third disk ends up beeing the second one in Linux (I’m not talking about device mapping) - you got the wrong mainboard. But one can leave with that because most mainboards found in cheap Windows desktop computers were not designed to boot several Linuxes. A bootmanager provides a way to handle the situation smoothly.

Have you got any more information sources about such limitations?

If you add a third hard disk in the third SATA port, set the BIOS order properly or not a all - so that it would just follow the order of the SATA ports, and the third disk ends up beeing the second one in Linux (I’m not talking about device mapping) - you got the wrong mainboard.

I hoped that my board “Asus M4N98TD EVO” (with BIOS 1101 from 2010-11-22) would not result in the reported surprise together with the third disk. :wink: