I just replaced a failed member of a raid1 volume and I can't get the OS (Gentoo) to recognize the new drive. I really can't reboot this server right now. Is there a way to poll the hard drives without rebooting?

thanks.

I assume we are talking about IDE here? IDE drives are not hot-swappable (aside from things like some laptop drive bays), which is a hardware problem and not something you can resolve in software.

SATA is another story though, and you shouldn't have any problems swapping them out.

You can make the system rescan the bus. How you do that depends on what bus you want to scan.

For scsi, execute this command:
echo "- - -" > /sys/class/scsi_host/host0/scan

I believe the same or very similar syntax works for rescanning a sata bus.

I don't believe you need to rescan an ide bus; just try to mount the device with its ID based on its position in the IDE chain.

Hi,
You can only hot swap IDE if the drive is mounted in a special tray that converts to a SATA;

This has been done but by the time you purchase the kit you might as well get the rack designed for Hotswap SATA and the drive kit for that. You wouldn't need to get the additional shockmount kit for the IDE.

Yes, you shouldn't hotswap IDE drives unless you have specifical designs to allow such.

That's what I needed to know. Thank you! I'll try that as soon as I get into the office.


BTW, they are hot swappable SATA drives. I should have been more specific in my original post.

OK. I tried the command and this is what I got:

echo "- - -" >/sys/class/scsi_host/host0/scan
bash: /sys/class/scsi_host/host0/scan: No such file or directory

The drives are definately hot swappable SATA drives, but the system identified them as /dev/hda and /dev/hdb. I've tried two different replacement drives and neither one is automatically recognized by the system. This is more complicated than I thought.

I inherited this server and there's no documentation, but I'm fairly certain the RAID volumes were created using mdadm. Here's the output from mdadm - D:

*********************
webdev / # mdadm -D /dev/md0
/dev/md0:
Version : 00.90.03
Creation Time : Thu Sep 29 07:19:53 2005
Raid Level : raid1
Array Size : 72192 (70.51 MiB 73.92 MB)
Used Dev Size : 72192 (70.51 MiB 73.92 MB)
Raid Devices : 2
Total Devices : 1
Preferred Minor : 0
Persistence : Superblock is persistent

Update Time : Wed Aug 19 14:55:47 2009
State : clean, degraded
Active Devices : 1
Working Devices : 1
Failed Devices : 0
Spare Devices : 0

UUID : 0c9e4f51:eeed1593:ad4f1fcc:48abf372
Events : 0.386

Number Major Minor RaidDevice State
0 0 0 0 removed
1 3 65 1 active sync /dev/hdb1
**************************************************

Here's the contents of /etc/raidtab:

***********************************
webdev / # cat /etc/raidtab
# /boot (RAID 1 MIRROR)
raiddev /dev/md0
raid-level 1
nr-raid-disks 2
chunk-size 32
persistent-superblock 1
device /dev/sda1
raid-disk 0
device /dev/sdb1
raid-disk 1

# / (RAID 1 MIRROR)
raiddev /dev/md2
raid-level 1
nr-raid-disks 2
chunk-size 32
persistent-superblock 1
device /dev/sda3
raid-disk 0
device /dev/sdb3
raid-disk 1
***********************************************

Fdisk -l yields:

**************************************************
Disk /dev/hdb: 74.3 GB, 74355769344 bytes
255 heads, 63 sectors/track, 9039 cylinders
Units = cylinders of 16065 * 512 = 8225280 bytes

Device Boot Start End Blocks Id System
/dev/hdb1 1 9 72261 fd Linux raid autodetect
/dev/hdb2 10 259 2008125 82 Linux swap / Solaris
/dev/hdb3 260 9039 70525350 fd Linux raid autodetect

Disk /dev/md0: 73 MB, 73924608 bytes
2 heads, 4 sectors/track, 18048 cylinders
Units = cylinders of 8 * 512 = 4096 bytes

Disk /dev/md0 doesn't contain a valid partition table

Disk /dev/md2: 72.2 GB, 72217853952 bytes
2 heads, 4 sectors/track, 17631312 cylinders
Units = cylinders of 8 * 512 = 4096 bytes

Disk /dev/md2 doesn't contain a valid partition table
*****************************************************

The SATA controller is:
IDE interface: Intel Corporation 82801EB (ICH5) SATA Controller (rev 02)


Right now, I'm afraid to even reboot this system because I'm not certain that hdb1 has an MBR or a valid partition table.

Any ideas? I'd be very grateful. Thanks.

take a look in /sys/class/scsi to see if you have anything.

I don't run sata here at all (All SCSI but for a handful of PATA) so I haven't encountered the particular problem.

The design and intentions of the /sys virtual filesystem is such that there has to be some appropriate entry for sata someplace in it.

I can't help you with mdadm; where I run RAID it is hardware raid using Adaptec SCSI controllers.

I don't have a /sys/class/scsi on this system, but there are several scsi directories. They're all empty.

webdev class # ls -lah scsi_*
scsi_changer:
total 0
drwxr-xr-x 2 root root 0 Dec 22 2008 .
drwxr-xr-x 37 root root 0 Dec 22 2008 ..

scsi_device:
total 0
drwxr-xr-x 2 root root 0 Dec 22 2008 .
drwxr-xr-x 37 root root 0 Dec 22 2008 ..

scsi_disk:
total 0
drwxr-xr-x 2 root root 0 Dec 22 2008 .
drwxr-xr-x 37 root root 0 Dec 22 2008 ..

scsi_generic:
total 0
drwxr-xr-x 2 root root 0 Dec 22 2008 .
drwxr-xr-x 37 root root 0 Dec 22 2008 ..

scsi_host:
total 0
drwxr-xr-x 2 root root 0 Dec 22 2008 .
drwxr-xr-x 37 root root 0 Dec 22 2008 ..

scsi_tape:
total 0
drwxr-xr-x 2 root root 0 Dec 22 2008 .
drwxr-xr-x 37 root root 0 Dec 22 2008 ..
*********************************************

I'm totally stumped.

Yah, I much prefer hardware raid and use it for the systems I build.