With Centos3, fedora1/2 I have been able to put /boot on a RAID1 partition and have it boot normally. Unlike earlier versions that used lilo as the default boot loader, these did not make the 2nd drive bootable but at least I was fairly confident that if the first drive failed I could use the install CD in rescue mode to fix up the 2nd one and go on. However, Centos4 won't boot at all with a similar install attempt and running grub-install in rescue mode gives an error about not being able to find the bios drives. Is there a howto somewhere to work around this problem?
My mainboard have SATA Raid function (sata_sil).
I want to create software raid on Centos 4, but I can't see mkraid packeage. How would I create software raid 0 on centos 4.
Thanks.
On Thu, 2005-04-07 at 17:33 +0700, Bank wrote:
My mainboard have SATA Raid function (sata_sil).
I want to create software raid on Centos 4, but I can't see mkraid packeage. How would I create software raid 0 on centos 4.
use the program:
mdadm
see "man mdadm" for details :)
I use dmraid as you see in. http://www.net-forums.net/forums/Mounting_Ntfs_Partions_On_Raid_0_Device-t19 791.html
my mainboard is MSI RS480 M2 http://www.msi.com.tw/program/products/mainboard/mbd/pro_mbd_detail.php?UID= 639 CPU AMD 64 3000+ HDD Segate 20G IDE Maxtor 160G SATA x 2
My mainboard have silicon image SATA Raid controller building. Then I compile kernel 2.6.11 and enable data-mapper module.
And I create Maxtor 160G x 2 to raid 0 . [root@ares ~]# dmraid -s *** Active Set name : sil_afaeafbgccbgb size : 640339968 stride : 64 type : striped status : ok subsets: 0 devs : 2 spares : 0 [root@ares ~]# sfdisk --show-size /dev/mapper/sil_afaeafbgccbgb 320169984 [root@ares ~]# sfdisk --show-size /dev/mapper/sil_afaeafbgccbgb1 320167386
Every thing seem to be OK, but when I mount /dev/mapper/sil_afaeafbgccbgb1 and use command df.
[root@ares ~]# df Filesystem 1K-blocks Used Available Use% Mounted on /dev/hda3 36384656 2596712 31939668 8% / /dev/hda1 101105 46194 49690 49% /boot none 504892 0 504892 0% /dev/shm /dev/mapper/sil_afaeafbgccbgb1 160074724 17954072 142120652 12% /home/topwarez.com/torrent/httpd/users
Why the system show a half size (160074724 ) of my raid array.
-----Original Message----- From: centos-bounces@centos.org [mailto:centos-bounces@centos.org] On Behalf Of Johnny Hughes Sent: Thursday, April 07, 2005 5:41 PM To: CentOS ML Subject: Re: [CentOS] How to make software raid on Centos 4
On Thu, 2005-04-07 at 17:33 +0700, Bank wrote:
My mainboard have SATA Raid function (sata_sil).
I want to create software raid on Centos 4, but I can't see mkraid
packeage.
How would I create software raid 0 on centos 4.
use the program:
mdadm
see "man mdadm" for details :)
I note that in the debian RAID is says to make two swap parititions of equal priority,
A better way is to layer swap on a proper mirrored raid device instead, this way the machine will still boot if one drive fails,
rather than complaining about missing swap...
P.
Bank wrote:
I use dmraid as you see in. http://www.net-forums.net/forums/Mounting_Ntfs_Partions_On_Raid_0_Device-t19 791.html
my mainboard is MSI RS480 M2 http://www.msi.com.tw/program/products/mainboard/mbd/pro_mbd_detail.php?UID= 639 CPU AMD 64 3000+ HDD Segate 20G IDE Maxtor 160G SATA x 2
My mainboard have silicon image SATA Raid controller building. Then I compile kernel 2.6.11 and enable data-mapper module.
And I create Maxtor 160G x 2 to raid 0 . [root@ares ~]# dmraid -s *** Active Set name : sil_afaeafbgccbgb size : 640339968 stride : 64 type : striped status : ok subsets: 0 devs : 2 spares : 0 [root@ares ~]# sfdisk --show-size /dev/mapper/sil_afaeafbgccbgb 320169984 [root@ares ~]# sfdisk --show-size /dev/mapper/sil_afaeafbgccbgb1 320167386
Every thing seem to be OK, but when I mount /dev/mapper/sil_afaeafbgccbgb1 and use command df.
[root@ares ~]# df Filesystem 1K-blocks Used Available Use% Mounted on /dev/hda3 36384656 2596712 31939668 8% / /dev/hda1 101105 46194 49690 49% /boot none 504892 0 504892 0% /dev/shm /dev/mapper/sil_afaeafbgccbgb1 160074724 17954072 142120652 12% /home/topwarez.com/torrent/httpd/users
Why the system show a half size (160074724 ) of my raid array.
-----Original Message----- From: centos-bounces@centos.org [mailto:centos-bounces@centos.org] On Behalf Of Johnny Hughes Sent: Thursday, April 07, 2005 5:41 PM To: CentOS ML Subject: Re: [CentOS] How to make software raid on Centos 4
On Thu, 2005-04-07 at 17:33 +0700, Bank wrote:
My mainboard have SATA Raid function (sata_sil).
I want to create software raid on Centos 4, but I can't see mkraid
packeage.
How would I create software raid 0 on centos 4.
use the program:
mdadm
see "man mdadm" for details :)
CentOS mailing list CentOS@centos.org http://lists.centos.org/mailman/listinfo/centos
Raid0? I am assuming you are going for the speed benefits(which are neglible at worst and not 2x the speed at best) instead of data security?
Bank wrote:
I use dmraid as you see in. http://www.net-forums.net/forums/Mounting_Ntfs_Partions_On_Raid_0_Device-t19 791.html
my mainboard is MSI RS480 M2 http://www.msi.com.tw/program/products/mainboard/mbd/pro_mbd_detail.php?UID= 639 CPU AMD 64 3000+ HDD Segate 20G IDE Maxtor 160G SATA x 2
My mainboard have silicon image SATA Raid controller building. Then I compile kernel 2.6.11 and enable data-mapper module.
And I create Maxtor 160G x 2 to raid 0 . [root@ares ~]# dmraid -s *** Active Set name : sil_afaeafbgccbgb size : 640339968 stride : 64 type : striped status : ok subsets: 0 devs : 2 spares : 0 [root@ares ~]# sfdisk --show-size /dev/mapper/sil_afaeafbgccbgb 320169984 [root@ares ~]# sfdisk --show-size /dev/mapper/sil_afaeafbgccbgb1 320167386
Every thing seem to be OK, but when I mount /dev/mapper/sil_afaeafbgccbgb1 and use command df.
[root@ares ~]# df Filesystem 1K-blocks Used Available Use% Mounted on /dev/hda3 36384656 2596712 31939668 8% / /dev/hda1 101105 46194 49690 49% /boot none 504892 0 504892 0% /dev/shm /dev/mapper/sil_afaeafbgccbgb1 160074724 17954072 142120652 12% /home/topwarez.com/torrent/httpd/users
Why the system show a half size (160074724 ) of my raid array.
-----Original Message----- From: centos-bounces@centos.org [mailto:centos-bounces@centos.org] On Behalf Of Johnny Hughes Sent: Thursday, April 07, 2005 5:41 PM To: CentOS ML Subject: Re: [CentOS] How to make software raid on Centos 4
On Thu, 2005-04-07 at 17:33 +0700, Bank wrote:
My mainboard have SATA Raid function (sata_sil).
I want to create software raid on Centos 4, but I can't see mkraid
packeage.
How would I create software raid 0 on centos 4.
use the program:
mdadm
see "man mdadm" for details :)
CentOS mailing list CentOS@centos.org http://lists.centos.org/mailman/listinfo/centos
William Warren wrote:
Raid0? I am assuming you are going for the speed benefits(which are neglible at worst and not 2x the speed at best) instead of data security?
Perhaps data security really isn't an issue for the poster. For example, I've got a rack of servers with identical content. Except for the logs, everything on any of the servers is really expendible (as long as they ALL don't fail...heh). So I mirror the partition containing the log data and periodically aggregate all the logs to somewhere "safe." As a result, I feel comfy using RAID 0 for the data partitions on those servers.
Cheers,
C
Oh i understand. That was not criticism..just pure curiosity.
Chris Mauritz wrote:
William Warren wrote:
Raid0? I am assuming you are going for the speed benefits(which are neglible at worst and not 2x the speed at best) instead of data security?
Perhaps data security really isn't an issue for the poster. For example, I've got a rack of servers with identical content. Except for the logs, everything on any of the servers is really expendible (as long as they ALL don't fail...heh). So I mirror the partition containing the log data and periodically aggregate all the logs to somewhere "safe." As a result, I feel comfy using RAID 0 for the data partitions on those servers.
Cheers,
C
CentOS mailing list CentOS@centos.org http://lists.centos.org/mailman/listinfo/centos
On Thu, 2005-04-07 at 17:33 +0700, Bank wrote:
My mainboard have SATA Raid function (sata_sil).
I want to create software raid on Centos 4, but I can't see mkraid packeage. How would I create software raid 0 on centos 4.
Thanks.
And here is a good howto http://unthought.net/Software-RAID.HOWTO/Software-RAID.HOWTO.html
On Thu, 2005-04-07 at 17:33 +0700, Bank wrote:
My mainboard have SATA Raid function (sata_sil).
I want to create software raid on Centos 4, but I can't see mkraid packeage. How would I create software raid 0 on centos 4.
Thanks.
Sorry for the third post in a couple minutes :P
I see where the RAID howto I posted uses the mkraid tool ... and your confusion. mdadm can be used instead to create RAID, without raidtools ... here are some examples:
http://www.linuxsa.org.au/mailing-list/2003-07/1270.html (debain howto ... but mdadm commands are the same)
http://xtronics.com/reference/SATA-RAID-Debian.htm (debain howto ... but the mdadm commands are the same)
Here is the mdadm author's page: http://cgi.cse.unsw.edu.au/~neilb/SoftRaid
Goggle has more mdadm examples...
On Tue, 05 Apr 2005 13:15:25 -0500, Les Mikesell lesmikesell@gmail.com wrote:
With Centos3, fedora1/2 I have been able to put /boot on a RAID1 partition and have it boot normally. Unlike earlier versions that used lilo as the default boot loader, these did not make the 2nd drive bootable but at least I was fairly confident that if the first drive failed I could use the install CD in rescue mode to fix up the 2nd one and go on. However, Centos4 won't boot at all with a similar install attempt and running grub-install in rescue mode gives an error about not being able to find the bios drives. Is there a howto somewhere to work around this problem?
This is an infamous grub-on-sw-raid problem. I run this script after installing the OS on the new system (and after grub upgrades, it used to fix flipping md5sums in Tripwire reports for me): #!/bin/sh # # Update MBR on both mirror drives # (grub/swraid has problems updating the slave, # as of RHEL3) # Should be run after grub rpm updates as well # # By Alex Tkachenko alex@ingrian.com
ADMINDIR=/root/admin
BOOT_ARRAY=`df /boot | awk '/dev/{print $1}'`
# Select only active disks (skip spares) DISKS=`mdadm --query --detail /dev/md0 | awk '/active sync/{print $7}'| sed ' s@/dev/@@g s/,/ /g s/[0-9]//g '`
for d in $DISKS do
cat <<EOF | /sbin/grub --batch device (hd0) /dev/$d root (hd0,0) setup (hd0) quit EOF
# Save updated mbr dd if=/dev/$d of=$ADMINDIR/mbr.$d count=1 done
Take care, Alex
-- Les Mikesell les@futuresource.com
CentOS mailing list CentOS@centos.org http://lists.centos.org/mailman/listinfo/centos
On Fri, 2005-04-08 at 00:37, Alex Tkachenko wrote:
However, Centos4 won't boot at all with a similar install attempt and running grub-install in rescue mode gives an error about not being able to find the bios drives. Is there a howto somewhere to work around this problem?
This is an infamous grub-on-sw-raid problem. I run this script after installing the OS on the new system (and after grub upgrades ...
Thanks - this is just the first time I'd seen where it didn't at least install on the first drive automatically. I had to run grub by hand from a rescue cd boot to fix it. Also, most of the instructions I found about installing on the 2nd drive were for IDE controllers and a bit different. Your approach gets it right for SCSI where a failure on the 1st drive makes the device names shift up. At least it works in a simulated failure where the 1st drive is removed. Does anyone know if there is a common failure mode where the 1st SCSI drive is still detected at boot as /dev/sda but doesn't work?
Les Mikesell wrote:
With Centos3, fedora1/2 I have been able to put /boot on a RAID1 partition and have it boot normally. Unlike earlier versions that used lilo as the default boot loader, these did not make the 2nd drive bootable but at least I was fairly confident that if the first drive failed I could use the install CD in rescue mode to fix up the 2nd one and go on. However, Centos4 won't boot at all with a similar install attempt and running grub-install in rescue mode gives an error about not being able to find the bios drives. Is there a howto somewhere to work around this problem?
If you are used to LILO, it still exist in CentOS4, but isn't the default boot loader. You can specify it during the installation, or you can remove grub and install LILO later on. It handles /boot on RAID1 automatically.
If you want to continue using Grub, this post from Fedora mailing list might be helpfull:
https://www.redhat.com/archives/fedora-list/2005-March/msg05935.html
-
Aleksandar Milivojevic -
Alex Tkachenko -
Bank -
Chris Mauritz -
Johnny Hughes -
Les Mikesell -
Peter Farrow -
William Warren