After installing Fedora 9 on my box (and making love to it!), today I tried to activate fake(or software!)RAID 1. It's an amazing journey filled with joy, sorrow and suspense. Below is a little big description of the journey which I wanted to share with you hoping that you might get help of it or can contribute to it to overcome the unresovled difficulties and quest.
I do have S975XBX2 Intel(tm) Workstation Board with two SATA 160.0 GB hard disks. Before activating RAID , I had installed Fedora 9 on one disk (keeping the another at deep sleep
) to try different features of Sulphur as it was my first Linux GNU/Linux experience.
First, I needed to activate the RAID support from BIOS. So, I went into BIOS by pressing F2 key during the start of machine boot.Then I went to "Advanced" tab where I found "Drive Configuration" menu item. Entering into "Drive Configuration" menu I found "Configure SATA As" item which can take any of the 3 values viz. IDE, RAID, AHCI. It was set to IDE, I changed it to RAID. Finaly I pressed F10 key to save the settings to reboot.
But the booting seemed to be failed. After showing nice messages of GRUB it threw following lines, -
Code:
Red Hat nash version 6.0.52 starting
Unabled to access resume device (/dev/sda2)
mount: could not find filesystem '/dev/root'
setuproot: moving /dev failed: No such file or directory
setuproot: error mounting /proc: No such file or directory
setuproot: error mounting /sys: No such file or directory
Mount failed for selinuxfs on /selinux: No such file or directory
switchroot: mount failed: No such file or directory
Then I recollected that I was supposed to create the RAID Valume using Intel(tm) Matrix Storage Manager and make a fresh Fedora installation.
So I restarted the machine, I was asked to press "Ctrl + I" (which I ignored earlier) to manage the RAID volume using "Intel(tm) Matrix Storage Manager".I did so and followed a pretty-user-friendly screen of "Intel(tm) Matrix Storage Manager option ROM v5.6.2.1002 ICH7R wRAID5" to create a RAID 1 volume comprising of the 2 hard disks. The final screen after RAID 1 creation held a list like this, -
Code:
RAID Volumnes:
ID NAME Level Strip Size Status Bootable
0 SRRAID1 RAID1(Mirror) N/A 149.1 GB Normal Yes
Physical Disks:
Port Drive Model Serial # Size Type/Status(Volume ID)
2 ST3160215AS 9RA6Q3NM 149.1 GB Member Disk (0)
3 ST3160215AS 9RA6Q0C8 149.1 GB Member Disk (0)
(I don't know why it is showing 149.1GB as the SATA disks are of 160 GB)
Then I exited the Matrix Storage Manager screen and booted the machine from Fedora 9 installation DVD and started installing Fedora. But right after selecting the "Language" and "Keyborad Layout" option I was splashed by the following confirmation message,
Code:
The partition table on device mapper/ isw-ceefeeedbj_SRRAID1 was unreadable.
To create new partitions it must be initialized, causing the loss of ALL DATA on the drive.
The operation will override any previous installation choices about which drives to ignore.
Would you like to initialize this drive, erasing ALL DATA?
I clicked "Yes" button as ALL MY DATA were already erased when I did created RAID volume using Intel(tm) Matrix Storage Manager. Then I created the custom partition layout which looked like this,
Code:
Drive Mount Point/ Type Format Size Start End
RAID/Volume
Hard Drives
|-/dev/mapper/isw_ceefeeedbj_SRRAID1
|-/dev/mapper/isw_ceefeeedbj_SRRAID1p1 /boot ext3 \/ 196 1 25
|-/dev/mapper/isw_ceefeeedbj_SRRAID1p2 swap \/ 2047 26 286
|-/dev/mapper/isw_ceefeeedbj_SRRAID1p3 / ext3 \/ 47975 287 6402
|-/dev/mapper/isw_ceefeeedbj_SRRAID1p4
|-/dev/mapper/isw_ceefeeedbj_SRRAID1p5 /var ext3 \/ 81917 6403 16845
|-/dev/mapper/isw_ceefeeedbj_SRRAID1p5 /home ext3 \/ 20481 16846 19456
This partition table got created and rest of the installation went similar as that for Non-RAID system. Only few different things caught my attention. Like the "Boot Loader Device" dialog box containg two options to select boot loader from, and BIOS drive order.
Code:
(0) Master Boot Recod(MBR) - /dev/mapper/iswceefeeedbjSRRAID1
( ) First Sector of boot partition - /dev/mapper/iswceefeeedbjSRRAID1p1
BIOS Drive Order:
First BIOS drive: mapper/isw_ceefeeedbj_SRRAID1 152618 MB Linux device mapper
After complete installation the machine booted without any problem. In order to check if the RAID system will be working at the time of disk failure I shut down the machine and unplugged one hard disk (on port 3) and started the machine again. During start I entered into Intel(tm) Matix Storage Manager (pressing Ctrl +I) and found the status of RAID Value as "Degraded". So, that means Matrix storage detected the unplugging of one hard disk. I exited Matrix Storage Manager to proceed further. But Linux failed to boot. It threw some ugly lines after showing our nice GRUB messages.
Code:
Red Hat nash version 6.0.52. starting
ahci 0000:04:00.0 :MV_AHCI HACK: port_map 1f->f
device_mapper: table 253:0:mirror: Device lookup failure
device_mapper: reload ioctl failed: No such device or address
device_mapper: table ioctl failed: No such device or address
device_mapper: deps ioctl failed: No such device or address
int[1]: segfault at 10 ip 0015c0fa sp bf81ba0c error 4 in libdevmapper.so.1.02[14e000+15000]
nash received SIGSEGV! backtrace(14):
/bin/nash(0x8053bc1)
[0x12e40c]
/usr/lib/libnash.so.6.0.52(nashDmDevGetName +0x5a)[0x13d1e6]
/usr/lib/libnash.so.6.0.52[0x1395bf]
/usr/lib/libnash.so.6.0.52[0x1396e9]
/usr/lib/libnash.so.6.0.52(nashBdevIterNext+0x109)[0x139b73]
/usr/lib/libnash.so.6.0.52[0x139e0f]
/usr/lib/libnash.so.6.0.52(nashFindFsByName+0x6e)[0x139f08]
/usr/lib/libnash.so.6.0.52(nashAGetPathBySpec+0xa5)[0x13a022]
/bin/nash(0x80414f49b)
/bin/nash[0x8053a2c]
/bin/nash[0x8054134]
/lib/libc.so.6(_libcstart_main+0xe6)[ox4bc5d6]
/bin/nash[0x804b141]
I tried to start the machine by Disabling RAID from BIOS, but got same result. I guessed when one hard disk failed the RAID 1 system will not boot until you attach a new/repaired hard disk. I may be wrong as I did not see any RAID system before trying on my linux box.
Anyway, I replugged the hard disk on port 3, enabled RAID on BIOS and restarted the machine.
On Intel(tm) Matrix Storage Manager I found that the status of RAID volume is marked as "Rebuild" in yellow. And message is being shown stating that Rebuilt will be handled by operating system.
Fedora got started nicely and everything seems to be normal. But I don't know if the RAID is rebuilt or not. I tried several option using dmraid but nothing seems to show any indication that RAID drives are not synced and the "Rebuild" status on Intel(tm) Matrix Storage Manager is still being shown. So, I guess it is working as I may need to wait for a real hard disk crash to test it.
Code:
dmraid --debug --verbose -ay
Any comment from anybody who had tried(and succeed or failed) RAID1 on Fedora with isw(Intel Software RAID) will be very helpful to me or any body trying to make RAID system, becasue manual on it seems to be lacking on internet.
Linear Mode
