software raid, kickstart and pxe installation

[Stefano Turolla <sturolla hq eso org>]

Hi everybody,

i am trying to install (with kickstart) a Dell Poweredge 1650 with two
18GB identical scsi disks.
I want the two disks mirrored via raid software and two swap partitions
( one per disk ).

I am using redhat 7.3 with pxe network installation that is working fine
with other architectures and without RAID.

I rebuild the initrd.img image using modules and kernel from
kernel-BOOT-2.4.18-18.7.x because of a problem with gigabit ethernet
drivers.

The ditribution i am using is the redhat 7.3 with all latest patches
that i am currently using to install and update other 80 hosts without
problems.
All the boot process is ok, aic7xxx, e1000, nfs and other modules are
loaded correctly, the the kickstart installation (with the correct
configuration files starts.

PROBLEM:
The partitions are created but the raid creation fails i think because
raid module is not loaded, can you help me ?
What i mainly don't understand is why loading of other modules like scsi
and nfs works while raid driver are not found.
Do they need to be in a particular directory in initrd.img and is there
any way to force to load them from a particular path?

thanks in advance
ciao

stefano



The error message i get is

"An error occurred trying to format md0. This problem is serious and the
install cannot continue"


On the kernel side (<alt> <f3>)) i have

modules to insert raid0 raid1 xor raid5 msdos jbd ext3 reiserfs jfs xfs
lvm-mod

module(s) to raid0 raid1 xor raid5 msdos jbd ext3 reiserfs jfs xfs
lvm-mod not found


on <alt><f4>

SCSI device sda: 35566478 512 byte hwdr sectors (18220 MB)
        sda: sda1 sda2
SCSI device sdb: 35566478 512 byte hwdr sectors (18220 MB)
        sdb: sdb1 sdb2

Adding swap.

.....
<6>md: bind <sda1,1>
<6>md: bind <sdb1,1>
<6>sda1's event counter: 0000000
<6>sdb1's event counter: 0000000
<3>md: md0 raid array is not clean, starting background reconstruction
<6>md: RAID level 1 does not need chunksize! Continuing anyway
<3> personality 3 is not loaded 
<6>md: unbind <sda1,1>
<6>md: export_rdev(sda1)
<6>md: unbind <sdb1,1>
<6>md: export_rdev(sda1)




This is my kickstart configuration file
-------------------------------------------
lang en_US

langsupport --default en_US 

text

network --bootproto static --ip 134.171.xx.xx --gateway 134.171.yy.yy
--netmask 255.255.255.0 --nameserver 134.171.zz.zz

nfs --server server.eso.org --dir /linux/redhat/7.3/os/i386

# select keyboard
keyboard us

# reset Master Boot Record
zerombr yes

# destroy all existing partitions on all disks
clearpart --all 

part swap --size=2048 --ondisk=sda
part swap --size=2048 --ondisk=sdb
part raid.01 --size=2000 --grow --ondisk=sda
part raid.02 --size=2000 --grow --ondisk=sdb
raid / --level=1 --fstype=ext2 --device=md0 raid.01 raid.02


install



Here is the list of modules added to initrd.img in modules/modules.cgz
---------------------------------------------
ls -l 2.4.18-18.7.xBOOT/
total 775
-rw-r--r--    1 root     root        36764 Dec 20 11:35 3c59x.o
-rw-r--r--    1 root     root       136441 Dec 20 11:36 aic7xxx.o
-rw-r--r--    1 root     root        67563 Dec 20 11:35 e1000.o
-rw-r--r--    1 root     root        27212 Dec 20 11:36 eepro100.o
-rw-r--r--    1 root     root         3412 Jan  7 14:50 linear.o
-rw-r--r--    1 root     root        68225 Dec 20 11:37 lockd.o
-rw-r--r--    1 root     root        53264 Jan  7 14:50 lvm-mod.o
-rw-r--r--    1 root     root        11756 Jan  7 14:50 multipath.o
-rw-r--r--    1 root     root       100398 Dec 20 11:39 nfs.o
-rw-r--r--    1 root     root         5672 Dec 20 11:36 raid0.o
-rw-r--r--    1 root     root        18016 Dec 20 11:36 raid1.o
-rw-r--r--    1 root     root        21980 Jan  7 14:50 raid5.o
-rw-r--r--    1 root     root        83588 Dec 20 11:38 scsi_mod.o
-rw-r--r--    1 root     root        17522 Dec 20 11:38 sd_mod.o
-rw-r--r--    1 root     root        20703 Dec 20 11:38 sr_mod.o
-rw-r--r--    1 root     root        88552 Dec 18 12:28 sunrpc.o
-rw-r--r--    1 root     root        10148 Jan  7 14:50 xor.o


module-info
---------------------------------
Version 0
aic7xxx
        scsi
        "Adaptec AHA-2740, 28xx, 29xx, 39xx"
3c59x
        eth
        "3Com 3c590/3c595/3c90x/3cx980"
e1000
        eth
        "Intel EtherExpress/1000 gigabit"


modules.dep
----------------------------------
aic7xxx: scsi_mod
usb-storage: scsi_mod
lockd: sunrpc
nfs: sunrpc lockd
sd_mod: scsi_mod
raid5: xor

pcitable
----------------------------------
0x9004  0x7893  "aic7xxx"   "Adaptec|AIC-789x"
0x9004  0x7894  "aic7xxx"   "Adaptec|AIC-789x"
0x9004  0x7895  "aic7xxx"   "Adaptec|AHA-2940U/UW / AHA-39xx /AIC-7895"
0x9004  0x7896  "aic7xxx"   "Adaptec|AIC-789x"
0x9004  0x7897  "aic7xxx"   "Adaptec|AIC-789x"
0x9004  0x8078  "aic7xxx"   "Adaptec|AIC-7880U"
0x9004  0x8178  "aic7xxx"   "Adaptec|AHA-2940U/UW/D / AIC-7881U"
........
0x8086  0x1000 "e1000" "Intel Corp.|82542 Gigabit Ethernet Controller"
0x8086  0x1001 "e1000" "Intel Corp.|82543GC Gigabit Ethernet Controller"
0x8086  0x1004 "e1000" "Intel Corp.|82543GC Gigabit Ethernet Controller"
0x8086  0x1008 "e1000" "Intel Corp.|82544EI Gigabit Ethernet Controller"
0x8086  0x1009 "e1000" "Intel Corp.|82544EI Gigabit Ethernet Controller"
0x8086  0x100c "e1000" "Intel Corp.|82544GC Gigabit Ethernet Controller"
0x8086  0x100d "e1000" "Intel Corp.|82544GC Gigabit Ethernet Controller"
0x8086  0x100e "e1000" "Intel Corp.|Dell Integrated Gigabit Controller"