New Drive - can't copy OS

[dkellogg3]

Hey all - need a bit of help here. I got a new HD (160 GB Seagate) to get a dumpster-dive recovered PII system up and running as a server.

Problem is that the old drive is going bad (maxblast labels the drive as "failing"), evident by numerous bad sectors. The seagate utility partitions and formats the drive fine (drives D and E), but hangs at the "preparing to copy" screen. I tried partion-to-partition in Ghost, but after the first bad sector was recognized, the speed went from 100 mb/min to about 5 mb/min. I then tried to image the partition to E:\, which I would then load to D:\, but the same drop in speed occured.

I used no switches when I used Ghost.

I'm no pro with computers or Ghost, but I can generally stumble my way through things eventually. Please help! Is this s-l-o-w transfer with Ghost normal? Should I do and -IA, -IB, -ID, or -IR switch? Which one?

Sorry - not a premature bump - I know someone is going to ask this... New drive is on a Promise Ultra 100tx2, with updated drivers and bios to support large drives

Thanks,
Don

 

[dkellogg3]

Bump.

ANY help would be appreciated.

Thanks

 

[JimPhelpsMI]

Hi, I have a 12 gig laptop drive that acts somewhat like that. It looks great after an FDisk, but format slows to a crawl at the first bad sector. That's at about 10 %. From then on it takes days to finish because all sectors from that point are bad. Format is testing bad sectors. The brain that wrote the program must have told it to try each sector about 300 times. I can MAXBLAST it. Maxblast ignores bad sectors and leaves them unmarked. Scandisk will run the same way as Format after that. Anyway, the drive is definitely BAD! Jim

 

[dkellogg3]

Bump... any help?

 

[Ausm]

Try low level formatting.

Ausm

 

[dkellogg3]

Sorry, I'm a wannabe geek - please explain low-level format. Also, the problem is not with the new (160GB) drive, I can format that just fine. The problem is with the old one, which has the bad sectors.

I tried once to use xxcopy, but had used the seagate software to partition and format the drive. I'm going to try to use fdisk and format and then xxcopy.

Thanks for the response.

 

[dkellogg3]

I'm going to re-attempt this this weekend. Any advice?

 

[VirtualLarry]

Ok, is this older failing drive, one that you want to recover existing data off of, or re-format clean and re-use it for another system?

If the first, then try using Ghost 2003, using -IR -fro. Raw sector-by-sector imaging, ignore bad sectors and continue. Your destination image file/disk, will have to be restored to a physical HD, and then logical data-recovery will have to be performed on the filesystem, as the bad sectors that were not copied, will leave "holes" with zero-filled sectors in the destination disk image.

If the second, then you should use the appropriate mfg's diagnostic boot floppy, to "write zeros" or "erase disk" (same thing, really). It will write zeros to all user data sectors, and automatically re-map bad sectors in the process. If the drive's spare-sector remap table is full - the drive is toast anyways, junk it. It's really not worth saving, honest.

As for the slowdowns that you experienced, those could be caused by several things. The drive could simply be internally re-trying to read/write sectors, if the drive were damaged. It could also have triggered the disabling of the drive's internal caching mechanisms, if a bad sector was detected and the remap table was full. Finally, and this is actually the more likely reason - Ghost normally requests a multiple-sector transfer, as they are the most performance-efficient. However, if there is an error, then Ghost has to re-try the request, using smaller and smaller numbers of sectors, until it gets down to making single-sector transfer requests instead. That is likely what happened. I'm not sure if Ghost "throttles back up" or not after a time of non-error transfers, I haven't experimented enough to test.

 

[dkellogg3]

I'm getting rid of that sucker - with the bad sectors, and the *screaming* whine it emits, as soon as I get the info off of it it's going in the trash.

As far as getting the info to the new drive, I have two partions set up on the new drive - one just larger than the source drive (~20GB), and the other the rest of the space on the drive (~130GB). If I do partition-to-image then image-to-partition, I would use the D:\ partition on the new drive as the temporary holding area. But couldn't I just do partition-to-partition (old drive to C:\ on the new drive)?

If it works using the switches you've indicated, how do I go about "logical data-recovery". Fortunately, the majority of the bad sectors occur at areas that have not data. (Symantec Disk Doctor told me this).

Would lowering (or eliminating) the compression when creating the image helpout at all?

Thanks for the response

 

[VirtualLarry]

Perhaps I made it a bit too complicated. You could try to do a file-by-file copy, partition to partition, directly from the old drive to a similar-sized partition on the new drive, and include the "-fro" switch. I guess the only reason that I generally do things the way that I do is so that I can do the data-recovery step manually afterwards, I don't really trust software to perform automated data-recovery.

In the above case, Ghost will most likely not copy *any* portion of a file that had a bad sector contained within it, and if that file happeded to be a directory file, likely none of the files referenced by that directory file would be copied over either. But it would be the simplest solution, with the least amount of work for the end-user, at the potential risk of losing more data, depending on where the bad sectors actually lay in the filesystem.

Compression shouldn't really make much of a difference, but when I do raw sector-by-sector imaging, to files, I generally also disable any compression, because that allows splicing together the image files manually if need be.