Hey everyone,
First of all i want to point out that this is my first post, and as such i want to apologize if the content(s)/questions are misplaced; if they are please let me know and i will move this post elsewhere
.
(**EDIT I also want to preface this by saying that i have done a lot of "Googling" and forum-sifting to try to resolve these errors. Although with that said, i would hate to be redundant in posting this thread, so if anyone can link to a longer-standing post regarding these issues please email it to me and i will link to it here or close this topic all together**)
Before describing the issues i've been encountering, here's some information about the hardware i'm using, in case any of the specifics may tip you guys off as to why i may be encountering some of these errors:
Computer/Server: SuperMicro SuperServer 5013C-T (a few years old)
- Pentium P4SCE Motherboard
- Intel E7210 Chipset
- Pheonix AwardBIOS (idk what version, but it is the same format as the BIOS referenced in Chapter 7 of the official manual ~
http://www.supermicro.com/manuals/superserver/1U/MNL-0720.pdf)
Other Specifics:
- While this server has 2 hotswappable hard drive bays that were previously set up to raid each drive, I am trying to install SME onto only 1 hard drive (since the second one went bad mechanically) and therefor with no raid.
* I suspect that this may be the crux of my issue as i'll explain below With that out of the way, my issue is this: After installing SME (smeserver-8.1-i386.iso) from a usb flash drive and disabling raid configuration (via "sme raid=0" at boot) and then rebooting from the hard disk, i either encounter a "GRUB Hard Disk Error", or the GRUB section of the boot phase remains blank. The latter looks something like this...
(txtxtxtxt)
Verifying DMI Pool Data ........................
Grub _
...where the "_" indicates the blinking cursor.
Since I know vaguely that GRUB is part of the bootloader and that i should be able to physically read and if need be rewrite the files, i rebooted the SuperServer from a live usb with Wary Puppy 5.5, mounted the sda1/bay1 hard drive, and found the Grub directory.
From here i opened the device.map file and found that it referenced
two hard disks; hd0 being the flashdrive that i installed SME from (which is no longer connected to the system) and hd1 being the bay1 had drive. The file looked something like this...
# This device map was generated by anaconda
(hd1) /dev/sda
(hd0) /dev/sdb
...soooo, thinking that this was indicative of a raid setup i proceeded to modify the file so that hd0 referenced the bay1 hard drive (removing the reference of hd1) and editing the grub.conf accordingly. However, since rebooting gave me the same error i suspect that i somehow miss-edited the files or, as i would assume is more probable, there is more to the GRUB boot process than just using the contents of those two files.
Anyway, that's all sort of unnecessary since i don't know enough about linux to manually configure a boot, but regardless my first question precedes all of that:
If there are two hard drive references in the grub.devices file, did Anaconda install SME with a raid1 configuration? If so, shouldn't the parameters "raid=0" have prevented that? As i mentioned earlier, the SuperServer was originally set up to raid two hard drives, so i then thought that perhaps something in the BIOS/CMOS was forcing any sort of OS installation to enable raid (from here i'm just speculating and throwing ideas out there - i can't say that i really know what i'm talking about that this point
). After rebooting again into the BIOS Setup I found nothing that suggested a forced raid setup, however, i did disable parallel ports and messed around with some other Advanced I/O and Bios settings.
Finally, and since nothing that i had done so far has made any changes, i rebooted into Wary Puppy and 'installed,' for a lack of a better term, Grub4Dos into the bootloader of the bay1 hard disk (while 'installing' it identified what i am assuming was the SME partition as "Linux (sda1:PBS)." It is also worthy to note that at this point i had restored the Grub directory to it's original state). Upon reboot i was able to force the server to boot from the linux partition of the hard drive where the boot process stopped at "kernel panic!!! - not syncing: Attempted to kill init! _". After doing a fresh re-install of SME ("sme raid=none" and encountering the same errors as before) and then re-installing Grub4Dos and booting from sda1, GRUB then attempted to find the sda1 partition on the empty floppy disk drive...
/facepalm.
With all of that said, i'm sorry for the massive wall of text, and if you're reading this then thank you for drudging through it
! I hope that some of you may have experienced a similar problem (**EDIT: and then you were able to fix it soon thereafter
**) or recognize what the issue is since i have exhausted myself trying to troubleshoot it myself. The main reason why i decided to post here is because when searching for solutions (via googling the errors word for word, etc) the posts and fixes themselves almost never applied to CentOS or SME, with the exception of official documentation on the wiki site (which only mentions how to disable raid config during setup).
Anyway, i want to express my thanks ahead of time for any and all input you guys can give me. At this point i feel like the only way to solve this on my own would be to educate myself on linux coding and GRUB syntax/coding, and because of my complete ignorance in that area it has made encountering these errors all the more frustrating and daunting
.
Thanks again!
7 Replies
1938 Views