Support for large block devices

[fallberg]

Hello,
I have a Drobo. I have earlier placed a wish for support for multiple LUNs in the Bubba|TWO kernel since the Drobo default have a LUN size of 2TB and I have four TB of data (so I get two volumes but the Bubba only "sees" the first).
I made a reformat of my Drobo with a LUN size of 8TB in order to solve the problem in wait for multiple LUN support but now I get

Code: Select all

scsi 130:0:0:0: Direct-Access     TRUSTED  Mass Storage     2.00 PQ: 0 ANSI: 5
sd 130:0:0:0: [sdf] Too big for this kernel. Use a kernel compiled with support for large block devices.

So apparently the kernel also lack support for large block devices.
Please add this support to the Bubba kernel since I can imagine many want to use their Bubbas for storage of data and hard-drive volume sizes increase rapidly. Since the Bubba now also supports LVM I guess this will become even more necessary.

Thank you!
BR, Patrick

[carl]

Hello,
I have a Drobo. I have earlier placed a wish for support for multiple LUNs in the Bubba|TWO kernel since the Drobo default have a LUN size of 2TB and I have four TB of data (so I get two volumes but the Bubba only "sees" the first).
I made a reformat of my Drobo with a LUN size of 8TB in order to solve the problem in wait for multiple LUN support but now I get

Code: Select all

scsi 130:0:0:0: Direct-Access     TRUSTED  Mass Storage     2.00 PQ: 0 ANSI: 5
sd 130:0:0:0: [sdf] Too big for this kernel. Use a kernel compiled with support for large block devices.

So apparently the kernel also lack support for large block devices.
Please add this support to the Bubba kernel since I can imagine many want to use their Bubbas for storage of data and hard-drive volume sizes increase rapidly. Since the Bubba now also supports LVM I guess this will become even more necessary.

Thank you!
BR, Patrick

The current kernel in claire (a 2.6.30 kernel) should have it activated.

[fallberg]

Ok. How do I use that kernel for my Bubba|TWO? I have preformed a software update through the admin interface so the system should be up-to-date.

BR, Patrick

[asparak]

You need to edit your sources file to enable the unstable repos, but I wouldn't recommend trying it unless you are confident you can fix things manually if it goes horribly wrong.

[fallberg]

Ah. Right. Thought I had already done that but I did a system restore some time ago due to my Bubba being bricked when I tried to update to the latest big change using the web-interface and then it was reset. I will try this.
Since things already have gone horribly wrong for me I think I can take that risk again now when I got the routine for recovery
I assume the update procedure will not put the USB-rescue feature out of business?

[tor]

Hi,

A quick comment here. Claire is our unstable repository. Its not guaranteed to give you even a functional system if upgrading against it. That said the kernel in it should, atm, be fairly stable for use on non critical systems. But only upgrade the bubba-kernel and bubba-modules from the repo.

fallberg, you should always be able to use the USB-revovery feature regardless on how badly shaped your hard drive is. It's a completely self contained "linux system" with all required parts.

/Tor

[fallberg]

Hi,
great. Then I have no worries. Trashed Linux systems does not scare me as long as there is a way to make the HW boot again (data saved or not).
I will test and report back on my success with the Drobo.

BR, Patrick

[fallberg]

I updated kernel and modules and it recognizes it properly according to dmesg. But I cannot mount the partition "by hand" and the web interface identifies the disk correctly (Mass Storage sdb 8796.1G TRUSTED) but the partition incorrectly (/dev/sdb1 Size: 2199.0G Failed to aquire suitable type of the filesystem on partition /dev/sdb1).
It seem that something is still missing for handling >2 TB volume sizes.

It mounts nicely on my Ubuntu 9.04 computer and the FS is ext3.

[tor]

Hi again fallberg,

What does mount and syslog say when trying to mount the partition?

/Tor

[fallberg]

Code: Select all

# mount -t ext3 /dev/sde1 /home/storage/extern/drobo
mount: wrong fs type, bad option, bad superblock on /dev/sde1,
       missing codepage or other error
       In some cases useful info is found in syslog - try
       dmesg | tail  or so

From syslog:

Code: Select all

Aug  7 20:31:46 bubba kernel: VFS: Can't find ext3 filesystem on dev sde1.

On my eeePC (running Ubuntu 9.04) 'sudo mount -t ext3 /dev/sdb1 /mnt' works just fine (yes the Drobo is identified as a different sd-device).

This is the dmesg output when the Drobo is inserted on the Bubba:

Code: Select all

sd 41:0:0:0: [sdd] Attached SCSI disk
scsi 42:0:0:0: Direct-Access     TRUSTED  Mass Storage     2.00 PQ: 0 ANSI: 5
sd 42:0:0:0: Attached scsi generic sg4 type 0
usb-storage: device scan complete
sd 42:0:0:0: [sde] Very big device. Trying to use READ CAPACITY(16).
sd 42:0:0:0: [sde] 17179869184 512-byte hardware sectors: (8.79 TB/8.00 TiB)
sd 42:0:0:0: [sde] Write Protect is off
sd 42:0:0:0: [sde] Mode Sense: 03 00 00 00
sd 42:0:0:0: [sde] Assuming drive cache: write through
sd 42:0:0:0: [sde] Very big device. Trying to use READ CAPACITY(16).
sd 42:0:0:0: [sde] Assuming drive cache: write through
 sde: sde1
sd 42:0:0:0: [sde] Attached SCSI disk

And this is the Ubuntu dmesg output (Ubuntu also automounts the volume, hence the EXT3 messages in the end):

Code: Select all

usb 1-2: new high speed USB device using ehci_hcd and address 6
usb 1-2: configuration #1 chosen from 1 choice
scsi5 : SCSI emulation for USB Mass Storage devices
usb-storage: device found at 6
usb-storage: waiting for device to settle before scanning
usb-storage: device scan complete
scsi 5:0:0:0: Direct-Access     TRUSTED  Mass Storage     2.00 PQ: 0 ANSI: 5
sd 5:0:0:0: Attached scsi generic sg4 type 0
sd 5:0:0:0: [sdb] Very big device. Trying to use READ CAPACITY(16).
sd 5:0:0:0: [sdb] 17179869184 512-byte hardware sectors: (8.79 TB/8.00 TiB)
sd 5:0:0:0: [sdb] Write Protect is off
sd 5:0:0:0: [sdb] Mode Sense: 03 00 00 00
sd 5:0:0:0: [sdb] Assuming drive cache: write through
sd 5:0:0:0: [sdb] Very big device. Trying to use READ CAPACITY(16).
sd 5:0:0:0: [sdb] 17179869184 512-byte hardware sectors: (8.79 TB/8.00 TiB)
sd 5:0:0:0: [sdb] Write Protect is off
sd 5:0:0:0: [sdb] Mode Sense: 03 00 00 00
sd 5:0:0:0: [sdb] Assuming drive cache: write through
 sdb: sdb1
sd 5:0:0:0: [sdb] Attached SCSI disk
sd 5:0:0:0: Attached scsi generic sg1 type 0
kjournald starting.  Commit interval 5 seconds
EXT3 FS on sdb1, internal journal
EXT3-fs: mounted filesystem with ordered data mode.

[tor]

Hi again,

Reading the output it seems like Bubba is fails to get the partition(sizes) of the drive thus reporing them as 2TB.

This could be due to Bubba, atm, does not support EFI GPT partition tables. Next 2.6.3x kernel will have this support and the testing kernel 2.6.26.5-9 should already have this. (The 2.6.3x is in our unstable repo, the 2.6.26.5-9 is in our testing repo which atm is targeted for our 1.2.1 update)

/Tor

[fallberg]

Hi,
ok. I am currently using this kernel:

Code: Select all

uname -a
Linux bubba 2.6.30.1 #4 Fri Jul 17 14:46:22 CEST 2009 ppc GNU/Linux

But perhaps that particular one does not have the appropriate flags enabled.

[tor]

Hi,

No unfortunately that is not enabled in that version. If you want to, it should be enabled in the kernel scheduled for our next bugfix upgrade 1.2.1, bubba-kernel_2.6.26.5-9 and matching bubba-modules.

The 2.6.30.x is our unstable kernel that we are working on for the 1.3 release.

/Tor

[fallberg]

Hi,
ok. How do I obtain the bubba-kernel_2.6.26.5-9 and modules? I would be happy to test out that functionality for you
I tried to modify the sources.list, but I do not know the explicit name of the packages. It will always fall back to 2.6.30.1

/Fallberg

[tor]

Hi again,

I think the easiest solution is to manually download the two packages from our repo, ie without using apt.

All versions of the bubba-kernel package can be found here:
http://update.excito.org/pool/main/b/bubba-kernel/

and the same url for the bubba-modules:
http://update.excito.org/pool/main/b/bubba-modules/

Download both packages to your bubba, become root and install by:

Code: Select all

dpkg -i bubba-kernel_xx.xx.xx_powerpc.deb bubba-modules_xx.xx.xx_powerpc.deb

Reboot your unit and it should be up and runnning the new kernel.

The other more correct solution is to change your sources lists to point to our testing repo, estelle, instead of the unstable claire repo.

/Tor