LinuxQuestions.org

LinuxQuestions.org (/questions/)
-   Linux - Server (https://www.linuxquestions.org/questions/linux-server-73/)
-   -   Kernal Error on Compaq Proliant SCSI Raid (https://www.linuxquestions.org/questions/linux-server-73/kernal-error-on-compaq-proliant-scsi-raid-628719/)

sixvoltsystem 03-17-2008 12:47 PM

Kernal Error on Compaq Proliant SCSI Raid
 
I'm pretty much a noob when it comes to linux, but I like diving into things. I have a Compaq Proliant 1850R that has 4 18.2 gig Compaq 10000 RPM drives in a Raid Config. I'm not sure what Raid config it is maybe 1 or 5 not sure. In grub it is showing Red Hat Enterprise Linux WS 2.4.21-4.
I obtained this server from my friends work (he owns it) and this server went down. There is just some personal word documents and one database he wants off it and like all things I get that are broken, have a strange fascination as to how it works and if can I fix it. Anyway my question to all you great people is:

#1: If this RAID Mirror was broken would it still load up grub?

#2: I get this error message when trying to load:

mkrootdev: Label / not found
Mounting root filesystem
mount: error 2 mounting ext3
pivotroot: pivot_root(sysroot,/sysroot/initrd) failed:2
unmount /initrd/proc failed: 2
Freeing unused kernal memory:224k freed
Kernal panic: No init found. Try passing init=option to kernal

Again I know I'm probly diving in over my head but you have to start somewhere and I find this most enjoyable.

As for the Hardrives themselves they are SCSI and they all show and sound to be spinning and working. Just odd. I don't feel like reinstalling Linux. Knoppix live cd would not load up I just get a blank screen. If I can fix it fine if not no big deal we'll send it to a data recovery place. But rather then shelling out tons of money I would like to try and fix it. Again thank you for any and all help.

TheDirtyScreech 03-17-2008 02:00 PM

Quote:

Originally Posted by sixvoltsystem (Post 3091666)
#1: If this RAID Mirror was broken would it still load up grub?

**The following assumes RAID is taken care of in software**

That depends on how you setup grub and RAID. To correctly have failover for RAID 1, grub should be installed on all drives, *and* installed such that every drive assumes it's the first drive. If it's installed on all drives, but the first drive fails, and each remaining drive thinks it's drive N (where N isn't the first drive), they won't actually load grub.

If you were running software RAID 5 on your boot partition, you shouldn't even correctly load grub, so I doubt that's the case.

If RAID is taken care of at the hardware level (and data isn't corrupted), one drive failure shouldn't ruin anything with either RAID 1 or 5. If more than one drive has failed, RAID 5 wouldn't load, and RAID 1 would depend on how many drives you had (3 or more before the two were corrupted and it would still load).



Quote:

#2: I get this error message when trying to load:

mkrootdev: Label / not found
Mounting root filesystem
mount: error 2 mounting ext3
pivotroot: pivot_root(sysroot,/sysroot/initrd) failed:2
unmount /initrd/proc failed: 2
Freeing unused kernal memory:224k freed
Kernal panic: No init found. Try passing init=option to kernal
First thing to check is that you have the correct filesystems (looks like ext3) compiled into the kernel (not a module). If that's the case, you will likely need to pass a "root=" or "real_root=" parameter in your grub config for the kernel. It should point to your "/" partition.

-TDS-

sixvoltsystem 03-17-2008 06:06 PM

Quote:

Originally Posted by TheDirtyScreech (Post 3091722)
**The following assumes RAID is taken care of in software**

That depends on how you setup grub and RAID. To correctly have failover for RAID 1, grub should be installed on all drives, *and* installed such that every drive assumes it's the first drive. If it's installed on all drives, but the first drive fails, and each remaining drive thinks it's drive N (where N isn't the first drive), they won't actually load grub.

If you were running software RAID 5 on your boot partition, you shouldn't even correctly load grub, so I doubt that's the case.

If RAID is taken care of at the hardware level (and data isn't corrupted), one drive failure shouldn't ruin anything with either RAID 1 or 5. If more than one drive has failed, RAID 5 wouldn't load, and RAID 1 would depend on how many drives you had (3 or more before the two were corrupted and it would still load).





First thing to check is that you have the correct filesystems (looks like ext3) compiled into the kernel (not a module). If that's the case, you will likely need to pass a "root=" or "real_root=" parameter in your grub config for the kernel. It should point to your "/" partition.

-TDS-

How would I go about doing that? Press "e" to edit the info during the grub menu? Sorry for being such a noob. Could you explain a little further about the process? Thank you so much for the help btw.

sixvoltsystem 03-17-2008 09:11 PM

Quote:

Originally Posted by TheDirtyScreech (Post 3091722)
**The following assumes RAID is taken care of in software**

That depends on how you setup grub and RAID. To correctly have failover for RAID 1, grub should be installed on all drives, *and* installed such that every drive assumes it's the first drive. If it's installed on all drives, but the first drive fails, and each remaining drive thinks it's drive N (where N isn't the first drive), they won't actually load grub.

If you were running software RAID 5 on your boot partition, you shouldn't even correctly load grub, so I doubt that's the case.

If RAID is taken care of at the hardware level (and data isn't corrupted), one drive failure shouldn't ruin anything with either RAID 1 or 5. If more than one drive has failed, RAID 5 wouldn't load, and RAID 1 would depend on how many drives you had (3 or more before the two were corrupted and it would still load).





First thing to check is that you have the correct filesystems (looks like ext3) compiled into the kernel (not a module). If that's the case, you will likely need to pass a "root=" or "real_root=" parameter in your grub config for the kernel. It should point to your "/" partition.

-TDS-

my grub config is:

root (hd0,0)
kernal /vmlinuz-2.4.21-4.ELsmp ro root=LABEL=/ 3
initrd /initrd-2.4.21-4.ELsmp.img

TheDirtyScreech 03-20-2008 02:59 PM

Sorry for the slow response, but work has been *really* busy lately.

Is this a previously working install or a completely new install? If it previously was working, what happened before it stopped working?

Quote:

How would I go about doing that? Press "e" to edit the info during the grub menu? Sorry for being such a noob. Could you explain a little further about the process?
Which part of my post are you referring to when you say "that?" "How would I go about doing..." what? Install grub on each drive such that each drive assumes it's the main drive? Check to see if ext3 is compiled into your kernel? Pass a "root=" or "real_root=" parameter in your grub config at bootup?


I need you to use pronouns as little as possible. I also want you to be posting as much information as you can. Often times, people will post only what they know will be necessary because they don't like typing. I'm not saying your doing that, but I want to make sure you don't do that.

Anyway, you are correct that you can (temporarily) edit your grub.conf by pressing the 'e' key.

Assuming your /boot partition is actually located on (hd0,0) and that your install of RH has the partition containing "/" actually labeled as "/" (RH does this labeling concept instead of just referring to the drive-partition mapping like the other Linux distros of the world), your grub.conf looks fine.

My guess is still that you don't have ext3 support compiled into the kernel, but have it compiled as a module instead. Did you compile your own kernel, or was it the RH default?

-TDS-

TheDirtyScreech 03-20-2008 03:03 PM

I just reread your first post and realized this isn't your server, but that it went down. Unless they compiled a new kernel, it isn't likely that the problem is ext3 not compiled into the kernel.

You may want to get your friend's information off before you continue. That can be done by connecting the drive to a working linux/bsd/*NIX system that has ext3 either compiled into the kernel or built as a loadable module. Then, mount the RH drive on a temp mount point and copy the info off.

If the RAID is a raid5 for the / partition, but either no RAID or RAID1 for the /boot partition, a drive failure could also show up like this. Can you ask your friend what level of RAID the machine was running?

-TDS-

sixvoltsystem 03-21-2008 05:22 PM

Quote:

Originally Posted by TheDirtyScreech (Post 3095254)
I just reread your first post and realized this isn't your server, but that it went down. Unless they compiled a new kernel, it isn't likely that the problem is ext3 not compiled into the kernel.

You may want to get your friend's information off before you continue. That can be done by connecting the drive to a working linux/bsd/*NIX system that has ext3 either compiled into the kernel or built as a loadable module. Then, mount the RH drive on a temp mount point and copy the info off.

If the RAID is a raid5 for the / partition, but either no RAID or RAID1 for the /boot partition, a drive failure could also show up like this. Can you ask your friend what level of RAID the machine was running?

-TDS-

First I'd like to say thank you, and sorry for the poor grammer. I appreciate everything you can do for me. Here is what happened:

#1. There are 4 drives total , each is a Compaq SCSI 10,000 RPM drive.

#2. This is his setup his previous IT guy setup. We have no installation cd's or info regarding what raid config it is.

#3. Its running RedHat Enterprise Edition...its an older version as for what one I have no idea, I am very very new to linux. My only experience is Fedora/Ubuntu, and open Suse.
#4. I may ask stupid questions so if your willing to put up with a complete idiot, I highly appreciate the help, if not I understand that as well.
#5. All he wants off the server is 2 files. But I can't seem to get any live cd to boot correctly. I don't have any access to hardware that would be able to handle ULTRA3 SCSI drives.

I think it is a RAID 1 config and I pray some of the info is retrievable
This all may be a lost cause, but things like this kill me when I can't figure them out. I almost went as far as to try to load up a EIDE hardrive on the CD-ROM port and install xp...then install fs-driver...copy the files then get out but I had problems trying that cause I didn't want grub getting written over. Any clues?

sixvoltsystem 03-21-2008 05:26 PM

One final note looks like the last drive is dying. All of them are spinning but the last drive shows a red hadrive light with an X through it.

Since I have access with grub might there be a way to view a drives contents within grub? I'm just glad its not lilo.


All times are GMT -5. The time now is 08:21 AM.