Adding and Using a New Disk

maccas17 · 05-01-2010, 09:36 AM

Hi,

I thought I had found a solution on the forum, but it doesn't seem to work - so I need a bit more help!

I'm running RHEL5:
2.6.18-164.el5 #1 SMP Tue Aug 18 15:51:54 EDT 2009 i686 i686 i386 GNU/Linux

under VMWare Server 2.0.2. When I create a new disk for my Linux VM I want to be able to use it without rebooting.

I checked what was currently in use:

cat /proc/scsi/scsi
Attached devices:
Host: scsi0 Channel: 00 Id: 00 Lun: 00
Vendor: VMware, Model: VMware Virtual S Rev: 1.0
Type: Direct-Access ANSI SCSI revision: 02
Host: scsi0 Channel: 00 Id: 01 Lun: 00
Vendor: VMware, Model: VMware Virtual S Rev: 1.0
Type: Direct-Access ANSI SCSI revision: 02
Host: scsi0 Channel: 00 Id: 02 Lun: 00
Vendor: VMware, Model: VMware Virtual S Rev: 1.0
Type: Direct-Access ANSI SCSI revision: 02

Also double-checked this with an ll /dev/disk/by-*

So I could see my new disk would be using HBA 0, channel 0, SCSI ID 2, and LUN 0. I would expect it to be /dev/sdd.

With the new 10MB disk created to the VM I issue:

echo "0 3 0" > /sys/class/scsi_host/host0/scan

This completes without error, but I do not have a /dev/sdd device file created for my disk.

Some dmesg output does seem to suggest the above command has worked and the disk has been found.

hdc: drive_cmd: status=0x51 { DriveReady SeekComplete Error }
hdc: drive_cmd: error=0x04 { AbortedCommand }
ide: failed opcode was: 0xec
Vendor: VMware, Model: VMware Virtual S Rev: 1.0
Type: Direct-Access ANSI SCSI revision: 02
target0:0:3: Beginning Domain Validation
target0:0:3: Domain Validation skipping write tests
target0:0:3: Ending Domain Validation
target0:0:3: FAST-40 WIDE SCSI 80.0 MB/s ST (25 ns, offset 127)
SCSI device sdd: 20480 512-byte hdwr sectors (10 MB)
sdd: Write Protect is off
sdd: Mode Sense: 5d 00 00 00
sdd: cache data unavailable
sdd: assuming drive cache: write through
SCSI device sdd: 20480 512-byte hdwr sectors (10 MB)
sdd: Write Protect is off
sdd: Mode Sense: 5d 00 00 00
sdd: cache data unavailable
sdd: assuming drive cache: write through
sdd: unknown partition table
sd 0:0:3:0: Attached scsi disk sdd
sd 0:0:3:0: Attached scsi generic sg3 type 0

If I reboot the system the disk is available and ready to use fine through the /dev/sdd device file.

Any ideas why the device file is not being created and how I can fix this?

Thanks.

smoker · 05-01-2010, 09:46 AM

I would think that disks are assigned when udev runs at boot, so maybe running start_udev as root would do the trick.

TBC Cosmo · 05-01-2010, 10:02 AM

Try doing a partprobe -s after you rescan the scsi bus. I'm not able to test myself.

maccas17 · 05-01-2010, 10:07 AM

Quote:

Originally Posted by smoker

I would think that disks are assigned when udev runs at boot, so maybe running start_udev as root would do the trick.

Thanks for the response.

I just created another disk for this VM to test your suggestion and this time the device file is created automatically and the disk is available to use

I'm more confused now!

I ran the same commands and saw the same output in syslog and dmesg.

Does anybody have experience of this not always working and if so, what to do (other than reboot!) in those circumstances?

Thanks.

maccas17 · 05-01-2010, 10:10 AM

Quote:

Originally Posted by TBC Cosmo

Try doing a partprobe -s after you rescan the scsi bus. I'm not able to test myself.

When I did the test fdisk -l (which I think is the equivalent) it did not list this disk. If there is no device file, I'm not sure tools such as these can attempt to read partition information.

TBC Cosmo · 05-01-2010, 10:20 AM

Quote:

Originally Posted by maccas17

When I did the test fdisk -l (which I think is the equivalent) it did not list this disk. If there is no device file, I'm not sure tools such as these can attempt to read partition information.

Fdisk provides the kernel's view of the partition table. Partprobe syncs the scsi midlayer changes witth the kernel. So no, they are not equivalent.

Since your new disk was visible to the scsi mid layer (/proc/scsi/scsi) it seemed to me that this could work. But it seems there is some flaky behavior from your description

maccas17 · 05-01-2010, 10:35 AM

Quote:

Originally Posted by TBC Cosmo

Fdisk provides the kernel's view of the partition table. Partprobe syncs the scsi midlayer changes witth the kernel. So no, they are not equivalent.

Since your new disk was visible to the scsi mid layer (/proc/scsi/scsi) it seemed to me that this could work. But it seems there is some flaky behavior from your description

Thanks for highlighting this difference to me; so I guess fdisk -l just reports the contents of /proc/partitions

Whilst, partprobe -s, as per the man page will:

"partprobe is a program that informs the operating system kernel of partition table changes, by requesting that the operating system re-read the partition table."

So from what you're saying and the man page info above the difference is that the OS will re-read disk partitions to re-populate the kernels view with the latest data - would this then have udev create device files for devices with no entries?

In other Unix flavours there are specific commands to create new device files for new hardware (aside from mknod), such as insf in HP-UX. Is there any equivalent for RHEL?

TBC Cosmo · 05-01-2010, 11:00 AM

I've had mixed results with this scsi-rescan script - attached. But it worked just now in esxi 4

Code:

[root@cent5-vm1 ~]# cat /proc/scsi/scsi
Attached devices:
Host: scsi0 Channel: 00 Id: 00 Lun: 00
  Vendor: VMware   Model: Virtual disk     Rev: 1.0 
  Type:   Direct-Access                    ANSI SCSI revision: 02
[root@cent5-vm1 ~]# fdisk -l

Disk /dev/sda: 5368 MB, 5368709120 bytes
255 heads, 63 sectors/track, 652 cylinders
Units = cylinders of 16065 * 512 = 8225280 bytes

   Device Boot      Start         End      Blocks   Id  System
/dev/sda1   *           1          13      104391   83  Linux
/dev/sda2              14         652     5132767+  8e  Linux LVM
[root@cent5-vm1 ~]# vim scsi_rescan.sh
[root@cent5-vm1 ~]# chmod 755 scsi_rescan.sh 
[root@cent5-vm1 ~]# ./scsi_rescan.sh 
Host adapter 0 (mptspi) found.
Scanning hosts  0 channels 0 for 
 SCSI target IDs  0 1 2 3 4 5 6 7 , LUNs  0
Scanning for device 0 0 0 0 ...
OLD: Host: scsi0 Channel: 00 Id: 00 Lun: 00
      Vendor: VMware   Model: Virtual disk     Rev: 1.0 
      Type:   Direct-Access                    ANSI SCSI revision: 02
Scanning for device 0 0 1 0 ...
NEW: Host: scsi0 Channel: 00 Id: 01 Lun: 00
      Vendor: VMware   Model: Virtual disk     Rev: 1.0 
      Type:   Direct-Access                    ANSI SCSI revision: 02
1 new device(s) found.               
0 device(s) removed.                 
[root@cent5-vm1 ~]# fdisk -l

Disk /dev/sda: 5368 MB, 5368709120 bytes
255 heads, 63 sectors/track, 652 cylinders
Units = cylinders of 16065 * 512 = 8225280 bytes

   Device Boot      Start         End      Blocks   Id  System
/dev/sda1   *           1          13      104391   83  Linux
/dev/sda2              14         652     5132767+  8e  Linux LVM

Disk /dev/sdb: 1073 MB, 1073741824 bytes
255 heads, 63 sectors/track, 130 cylinders
Units = cylinders of 16065 * 512 = 8225280 bytes

Disk /dev/sdb doesn't contain a valid partition table