Hallo
I submitted this as a bug several weeks ago, but I wanted to ask around & see if anyone else has come across this....
I have a USB Buffalo Drivestation Quattro, with 4 1TB disks configured in raid5 as one 2.8TB (or so) disk, attached to a Cent 5.4 64 bit server (completely yum'd up to date)
The disk is labeled as GPT, and formatted as a 2.8 TB ext3 partition (this issue also happens with xfs). I used a gparted boot disk to create the partition.
When I attach the drive I see this in messages:
Nov 2 14:26:55 kernel: usb 1-5.2: new high speed USB device using ehci_hcd and address 7 Nov 2 14:26:56 kernel: usb 1-5.2: configuration 0000001 http://bugs.centos.org/view.php?id=1 chosen from 1 choice Nov 2 14:26:56 kernel: scsi5 : SCSI emulation for USB Mass Storage devices Nov 2 14:26:56 kernel: usb-storage: device found at 7 Nov 2 14:26:56 kernel: usb-storage: waiting for device to settle before scanning Nov 2 14:27:01 kernel: Vendor: BUFFALO Model: HD-QSSU2/R5 1 Rev: 2.02 Nov 2 14:27:01 kernel: Type: Direct-Access ANSI SCSI revision: 05 Nov 2 14:27:01 kernel: sdc : very big device. try to use READ CAPACITY(16). Nov 2 14:27:01 kernel: sdc : READ CAPACITY(16) failed. Nov 2 14:27:01 kernel: sdc : status=0, message=00, host=5, driver=00 Nov 2 14:27:01 kernel: sdc : use 0xffffffff as device size Nov 2 14:27:01 kernel: SCSI device sdc: 4294967296 512-byte hdwr sectors (2199023 MB) Nov 2 14:27:01 kernel: sdc: Write Protect is off
After this failure, the disk is either a) inaccessible, or b) reports only a 2 TB partition.
The latest Ubuntu can read the disk, presenting the full 2.8 TB just peachy.
This server is up to date: uname -a: Linux myserver.mydomain.com 2.6.18-164.6.1.el5 #1 SMP Tue Nov 3 16:18:27 EST 2009 i686 i686 i386 GNU/Linux cat /etc/redhat-release: CentOS release 5.4 (Final)
[root@myserver ~]# cat /proc/partitions major minor #blocks name ... 8 32 2147483648 sdc << the disk showing incorrectly with only 2TB of storage
This bug seems very similar to a previous bug: https://bugzilla.redhat.com/show_bug.cgi?id=502944 which was reported fixed in 5.4
Anyone seen this before, or have any ideas how I can get CentOS to see the disk?
Cheers, Gareth
This bug seems very similar to a previous bug: https://bugzilla.redhat.com/show_bug.cgi?id=502944 which was reported fixed in 5.4
Anyone seen this before, or have any ideas how I can get CentOS to see the disk?
Centos is usually using "old" kernel, so if you want avoid this then possible you need to recompile your own kernel or patch current.
Bugs on usb-disk are too common nowadays..
-- Eero
Gareth Tupper wrote, On 11/16/2009 03:12 PM:
Hallo
I submitted this as a bug several weeks ago, but I wanted to ask around & see if anyone else has come across this....
what BZ and #? (mainly out of curiosity, but not enough to override the laziness of not wanting to check 2 different BZs)
I have a USB Buffalo Drivestation Quattro, with 4 1TB disks configured in raid5 as one 2.8TB (or so) disk, attached to a Cent 5.4 64 bit server (completely yum'd up to date)
<SNIP>
After this failure, the disk is either a) inaccessible, or b) reports only a 2 TB partition.
<SNIP>
[root@myserver ~]# cat /proc/partitions major minor #blocks name ... 8 32 2147483648 sdc << the disk showing incorrectly with only 2TB of storage
This bug seems very similar to a previous bug: https://bugzilla.redhat.com/show_bug.cgi?id=502944 which was reported fixed in 5.4
Anyone seen this before, or have any ideas how I can get CentOS to see the disk?
ideas: A1) figure out how much more/less than http://www.linuxhq.com/kernel/v2.6/24/drivers/usb/storage/usb.c needs patched into the kernel source to make >2TB work. A2) get the CentOS kernel SRPM and patch it in, build, install and use. [considering the bz you point to points to (in Comment #7) a very small patch for the ipbvscsi devices, it is _probably_ just a simple patch from the 24 version of usb.c]
B1) give a kernel dev at that prominent North American Enterprise Linux vendor (who runs Enterprise Linux instead of Fedora) a 2.8TB USB disk to play with and B2) point them at http://www.linuxhq.com/kernel/v2.6/24/drivers/usb/storage/usb.c
:)
Alternatively we could find someone with a 2+TB USB disk and the ability to submit bugs on a subscription to that prominent North American Enterprise Linux vendor. (or see if a proven change could be put in a CentOS plus kernel[module])