Plusieurs personnes sur ce forum (y compris moi) se sont déjà demandé s'il était possible de réparer un raid brocken, jusque là les réponses était à tendance négative, mais apparement il y a des possibilités.
Voici un lien (english) :
http://www.viahardware.com/faq/kg7 [...] roller.htm
Je fais un copier/coller car ce genre d'info est qq fois très volatile :
The Highpoint Controller is reporting a broken stripe set. How can I fix it?
Note that running the disk benchmark program HDTach with Advanced Size Check enabled is a frequent cause of broken stripe sets! Another likely cause of broken strip sets is a defective IDE ribbon cable - be sure to replace your ribbon cables if you suffer from a broken stripe. Some people have had particular problems with rounded IDE cables (although not all: I use them without problems).
The simplest thing to try first is to clear your CMOS and/or reflash the BIOS. This has been reported to fix reports of broken stripes in some circumstances. If this doesn't help, fear not! There are several means of repairing your disks. I advise that you try them in the order given here, as once you try one method there is a danger it will prevent other methods from working.
Users of the NTFS file system may also want to read Microsoft's knowledge base articles on How to Recover From a Corrupt NTFS Boot Sector and Recovering NTFS boot sector on NTFS partitions and Description of the Windows Recovery Console.
Use of Highpoint's raidrb utility
Highpoint technical support kindly sent me an early version of their new utility to fix broken stripes on their HPT370//372A controllers. In this early version of the tool it is only possible to fix a RAID-0 array comprising two disks, or a RAID 0+1 array. It will not repair four-disk RAID-0 arrays. The utility raidrb is available on the downloads page. To use the utility do the following:
Remove all IDE devices apart from the broken RAID array
Copy the utility onto a DOS floppy disk, boot from floppy into DOS and then run the utility by typing "raidrb" at the A: prompt (without quotes). You will be presented with four options.
For the first attempt at rebuilding the RAID array, select option 1 "Create the bak file.And then rebuild!". This will save the configuration of your RAID array and then attempt to rebuild the array.
For any subsequent attempts at rebuilding the RAID array, select option 2 "Don't create the bak file.And rebuild!". This will rebuild the array using a previously saved bak file.
If the rebuild fails, use option 3 "Resume these disks' information!" to restore the original disk configuration. This uses the bak file created when you ran option 1.
After running the utility, reboot, and see if you can access the disk from DOS (eg. type DIR). If you can, your disk is repaired!
If the utility fails to rebuild the array, Highpoint recommend you contact their support department. Alternatively you can try one of the other methods below - but be aware that these may prevent Highpoint from offering you an alternative solution to the above. Your choice.
Alternative approach using mrecover2
Another user reported a slightly simpler restoration method, using the utility mrecover2 (available on the downloads page). If the Highpoint controller reports a broken stripe then simply:
Delete the RAID-0 array in the Highpoint BIOS
Recreate the RAID-0 array using identical settings
Boot to DOS from floppy
Run mrecover2 to rebuild the partition and file access tables
Reboot
Alternative approach using Norton Disk Doctor for DOS
Delete the broken RAID-0 array in the Highpoint BIOS
Recreate the RAID-0 array in the BIOS using identical settings
Boot to DOS from a floppy also containing Norton Disk Doctor for DOS
Run Norton Disk Doctor and and let it rebuild all the partitions and file allocation tables (FATs). Answer "yes" to the various questions asked.
Alternative approach using Symantec's Diskedit and Norton Disk Doctor
Diskedit and Norton Disk Doctor can be found on the Symantec Systemworks 2002 CD (these only support FAT16 and FAT32 - not NTFS). The steps to take are as follows:
Use Diskedit to view absolute sector 0 on the first of the HD
on the broken array as a Partition Table sector. Record the numbers shown.
Delete the broken RAID-0 array in the Highpoint BIOS
Recreate the RAID-0 array in the BIOS using identical settings
Use Diskedit to re-write the sector 0 numbers found in step 1
Use Norton Disk Doctor to put the media descriptor back to F8, then FDISK and MBR to make it boot again.
Under WindowsXP use the repair facility of the installation disc to run CHKDSK, FIXBOOT and FIXMBR. Similar tools are available for other Windows versions.
Alternative approach using testdisk
If that fails, the following solution is very successful for most people. Many thanks to Yasin Abbas (YaZ) for this - the solution is all his: I've just altered the wording here and there! In his case, he accidentally broke the stripe configuration by connecting the Secondary striped hard disk in a different IDE configuration. Upon booting the Highpoint Controller declared the stripe broken. Even after placing the hard disks back into their old configuration the to Highpoint controller continued to say that the stripe was broken. The first hard disk was indicated as a broken stripe Array#0; the second was indicated as HDD1 (standalone). The following solution has worked for him on two occasions, but is quite involved. It may not work under all circumstances but it's certainly worth a try. Here goes.
Ensure disks are correctly installed. Place the hard disks into the configuration you want, making sure that they both remained in the same orientation. By orientation we mean that the first hard drive in the broken stripe remains as the first hard drive in the "to be fixed" stripe regardless of where it is on the IDE channel and the second hard disk in the stripe remains as the second hard disk in the "to be fixed" stripe, regardless of where it is on the IDE channel.
Delete broken stripeset and create a new stripeset. Now enter the Highpoint BIOS and set up your stripe again, by deleting the old broken one and starting again from scratch. The BIOS will tell you that it will erase your drives, don't worry, let it do so. You have nothing to lose - your data is currently inaccessible anyway!
Analyse the partitions using testdisk. Once it has created the new stripe set, boot into DOS. You now require testdisk (Yasin used version 3.4) - an excellent program written by Christophe Grenier available from his site at http://www.cgsecurity.org/testdisk.html or on the downloads page. Run testdisk and you will be presented with a list of logical hard disks, the CHS values and the size. Usually your stripe set will be shown as "Enhanced BIOS mode". Select the stripe set that you created and choose to "Analyse" the partitions.
MBR new stripe set. Sometimes stripe corruption will remove the Master Boot Record (MBR) and testdisk will give you an error. To replace the MBR run: fdisk /MBR or better still use bootpart (Version 2.20) another highly recommended freeware utility by Gilles Vollant to replace the MBR. Documentation on how to do this is provided with bootpart.
Use testdisk to recreate the partitions. Upon completion of the analysis if your original stripe set wasn't too garbled you'll be presented with a list of your old partitions and their various values and types. You can't do anything with this list, just select your only option "Quit". The next screen shows you the list of partitions, structures and their types. In this list you are able to set-up each available partition as one of 5 different types: Primary bootable, Primary, Logical, Extended or even deleted. At this point try to make sure you set things up as they were before. This screen may not appear on very badly damaged stripe sets. After doing this select "Ok". The next screen allows you to arrange partition ordering. The usual order is Primary bootable = 1, Extended = 2, etc (assuming you have no other primary partitions or any other type, so this may differ in your case). Right and left selects different orders, up and down different partitions. I don't think the ordering actually matters too much as long as each partition is ordered consecutively (1,2,3 . . .) and no partitions have the same order (1, 1, 2, 3, 3, . . . but I haven't tested this myself!). After doing this select "Ok". This is the crucial stage, the point where; if you've set things up correctly you should get access to your data. A list is shown with the modifications. Your only two options are "Quit" and "Write". At this point in time if you think you have chosen your previous ordering and other values correctly, then select "Write". Otherwise choose to "Quit" and restart the procedures again.
Copy accessible data to another drive. Upon writing the partitions hopefully you can access your data. Usually your partition tables are still garbled. YOU MUST COPY ALL YOUR DATA OFF OF THE STRIPE SET AND REPARTITION YOUR HARD DISK. Although your data is now accessible, your disk is still in a very bad state and cannot be relied on!! The partition tables are only given "best" values and not "actual" values. However, now that you can access your data you have the chance to backup your essentials (or everything) and rebuild you RAID array from the beginning.
Rebuild RAID array and restore data. Once you have accessed all your data and copied everything off of the broken stripe set you can safely erase your hard disks and start from scratch, create your RAID-0 array again and then create your partitions. (Yasin recommends Partition Magic for this, or you could use FDISK or freeware tools like FreeDOS FDISK or Ranish Partition Manager . If you know of any other good freeware partition manager please email me!)
raidrb peut-être téléchargé ici : http://download.viahardware.com/fa [...] raidrb.zip
mrecover2 peut-être téléchargé ici :
http://download.viahardware.com/fa [...] ecovr2.zip
Message édité par Armaniz le 22-08-2002 à 14:27:06
---------------
La réalité est ce que l'on perçoit d'elle. /// Config. /// Montre. /// Partage Smartphone/PC, via internet