array not bootable ! ( Pro 5000+ smartarray 2/P)

[kcpant]

hi friends,

one of my clients have one proliant 5000 server, smartarray 2/P installed and 3 9.1 GB UW SCSI HDD configured in RAID 5. one of my colligue serviced the server because of some dust problem. after that,the server is not able to boot! then I have cheked it through SCU & ACU, it shows one logical drive in RAID 5 (in proper condition, without any error).but when it was examined through O/S CD, the logical drive is seen as unpartitioned space, it means all the drive partitioning has wiped out.
so, can anybody tell me how to recover the array configuration, as well as the data partitions?. please also let me know the possibilities because of which this disaster happened.

thanks in advance!

[Rahul Sharma]

I am assuming that the other array was a raid 0+1 and had os on that.

still it is better to know before hand what your friend did. did he take outall the drivers and the even raid card from the server?

because if say for example in live environment the os hard drives go bad still the array config is still on raid controller card and if the RC card is bad still the info remains on the hard drive in either case one could be replaced and the information can be copied back to card of drives depending upon the situation.
you have to tell us what is exact error you are getting on the console.
I hope that explains.

[Rahul Sharma]

One more thing forgot to mention. may be boot option was changed or has to be reset.
you need to select the boot controller.

[SAKET_5]

Hi kc,

more more info please...

what do you see when you boot the server, at what stage does it die? could you still get into the RBSU and make sure that the OS selection is correct and the boot controller is correct?

do you see any lights on the drives? do you have a test server handy, in the worst scenario, you could move the array to a functional server and see if that isolates a motherboard fault or something?

what would explain the disaster....going back to the basics...was he using antistatic wrist straps, etc...?anything silly could have made the whole thing go south!!

good luck...

[kcpant]

friends,

I'm sorry to say but you haven't gone properly through the details I've mentioned above. I've already said that it's a RAID 5 volume, but one gentleman "assuming" it RAID 0+1. dear saket, I've already done the all things what could be done to resolve it. the logical volume ( and so the physical HDDs) is looking proper, I've checked it through ACU and ADU . the boot sequence is proper ( boot controller is set to array card in SCU).when I check the logical volume partition information, it shows no partitions defined ( all space is unpartitioned).it means, the array configuration data has been wiped out from both ( array card & HDDs) locations ( as I think from this situation).

So, though I'm 98% sure that data recovery is not possible in this situation, can anybody suggest me if there is any possibility to get it back working?

[Brian_Murdoch]

Hi,

Are you able to boot into the system partition (diagnostics etc.) by pressing F10 after POST? If so then it might suggest that only the O/S boot partition has been corrupted.

If this system was running NT4 then there is something more worth trying. If the Master Boot Record (MBR) becomes corrupt then the usual symptoms are that the system will go through POST and begin the boot procedure but the screen will remain black with only a flashing cursor in the top left side of the screen (On older NON-RBSU systems like the PL5000).

If the MBR is corrupt then it can be bypassed with an NT boot floppy.

Go to another NT4 (or 2000) system (Preferably another Compaq server) and format a floppy disk with FAT.
Copy NTLDR, NTDETECT.COM and BOOT.INI from the other system to the floppy.
If the Proliant 5000 had a system partition (F10), then this would be partition 1, so the operating system would be on partition 2. Check and edit BOOT.INI to make sure that it reflects the location of the O/S. For example on an older Proliant with a system partition, NT4 would normally be pointed to as follows in BOOT.INI :-

multi(0)disk(0)rdisk(0)partition(2)\WINNT="Windows NT Server Version 4.00"

Take this floppy and boot the Proliant 5000 with it. If the system comes up then you know that only the MBR is corrupt on the server and there is a fix. I will await your update.

Good luck.

Brian

[kcpant]

hi brian,


I'm not able to boot into system partition ( if there is no partition information, the system partition has also been gone out, which is the situation at my end).

the solution you have told about NT 4.0 , I 've alreasy tried it as my first step towards troubleshooting.

[e4services]

WOW! I hate when that happens.
Seriously, doesn't sound good. You say that the ACU shows the one logical drive, but then you ask how to recover the array configuration. I think that is where the other response assumed something was missing in the array configuration. So why do you ask to recover the array configuration if ACU shows the correct config?
If what you see is not correct, I would try to remove the drives and replace them in 3 other slots or another Proliant and see f you can get that controller to read a different array config off of the disks. Or at worst case remove the drives and clear the controller and then replace the drives to force the controller to read the config back in from the disks.

Strange that his services could cause such a loss. What even could happened to lose the partition information, but not the array information.
It is almost as if the exist array configuration was cleared and a new on created. Or the drive was formatted.
Did he remove the drives may be? replaced them in a different position or order? Guess that would not be true for you would still see the old array 'failed' in the ACU.

Anyway, I looks as if all is lost unless you can get the config back the way it was.

[Gus Kwong]

Hi Kcpant,

I had the same problem in the past (SmartArray's RAID5 logical drives all healthy and no error at all) and it was not recoverable for me --- if you cannot see any partitions information via Win2000 commandline boot from CD, then it is not looking good. You can try the standard Win2000 tools from the boot CD, like fixboot and fixmbr --- but don't hope too much because all your partition info not longer readable.

I would agree that this problem is not to do with Logical Drives --- it is more to do with Win2000.

I vaguely remember my problem was to do with corrupted partitions after upgrading to Dynamic Disks (but failed), and when Win2000 boot and access the hard-drives, it would access the disk successfully when it is using naming convention such as "Device(0)" but it will just fail during bootup when accessing the disks as "Partition(0)". I cannot remember exactly what, but you may have better idea if you check the mapping for your disks and partition (type HELP in cmd booted from CD for more details).

But there is still hope --- If data is important on the corrupted partitions, then you may need to consider about data recovery to bring back the partitions and even RAID5 disks... Let me emphasise this, you can recover *anything* (as long as you don't write anymore stuff on disks) even you reformat the disks if you go with the data recovery route.

Hope this helps.

Gus

[kcpant]

hi friends,


Thanks for replying to my query. You all are guessing right ( which is unfortunete to me!) that all data's been wiped out. I'm now following the data recovery companies to get the whole data back.

thanks again ..

regards