WEIRD MacPro softRAID problem HELP

I'm having a nearly identical situation with the same Maxtor MaxLine Pro drives you are using. I hadn't thought of Toast being a problem but the first time this happened to me I was also working in Toast 8 trying to burn additional copies from a VIDEO-TS folder. I was working off the internal raid with no additional firewire drives connected.

At first, the drives would dismount shortly after startup like they are a firewire or USB device. I've received the same error messages you've received.

Disk Utility won't fix it, indicating one drive has failed and the other is offline. In my case, it alternates between the two drives as to which has failed and which is offline, never the same answer twice in a row.

DiskWarrior rebuilt the directory a couple of times, and it's hardware test indicated there were no problems (as did Apple Hardware Test), but they would continue to dismount upon restart into OSX. Now DiskWarrior has stopped working and is indicating damage to the drives.

There have been no clicking or weird sounds coming from the drive so I suspect the directory is just hosed beyond repair. I don't know if the drive is salvagable and I'm uncertain at this point whether any of the data recovery software solutions will work on a RAID-0 striped pair. If not, we may be really screwed.

Have you simply tried removing the drives and reseating them and even swapping the order of the drives in the drive bays? Reordering the bays will do nothing to impact your RAID apart from possibly showing a potential issue with your MPs SATA bus.

What's interesting is that both of you were using Toast 8 when you both experienced the exact same issues. Makes me sort of glad to be still using Toast 7.

I know that I haven't used Toast since this started happening and the problems have escalated.

Haven't taken the drives out and reseated them yet though. I'm having to finish a deadline project on my old computer before I can troubleshoot further.

If I had a 4-drive RAID and 500GB not backed up, I'd be using SoftRAID 3.6
http://www.softraid.com

... which is what - by your title - I thought you were using,

and then use SynchronizePro! for scheduled backups, could be couple times a day to external SATA or Firewire. Maybe a hot-swap two drive enclosure (using 500GB drives like big floppies).

And Disk Warrior 4.

I have been using SoftRAID since 2.2.2 and OS 9.1 and it is really been rock solid.

Once and "if" it gets all the volumes and data back online, try and pull and backup the latest changes and such and then rebuild the array.

SR is good at notifying when there is an unrecoverable error so you know immediately when it happens. Or even recoverable errors.

My own experience, if using SoftRAID, is to not install a half dozen other utilities to monitor drives and smart status and just let it do its own thing.

I have seen the exact same problem happen twice now on my Mac Pro. Here is the configuration:

Mac Pro 3 GHz, 8 GB Apple RAM, nVidia FX 4500
5 internal SATA drives (all WD1500ADFD) one drive (optical bay) is used for system, 4 others are used in Apple RAID 0 (stripe) for data.

Two of the RAID failures happened while my backup software was running a backup script (Personal Backup X4) that clones the RAID to external fw800 drive. These failures reported the following in the system log:

kernel[0]: AppleRAID::completeRAIDRequest - error 0xe0030005 detected for set "Stripe" (0D6F397D-04AB-4D12-A634-73E668D9B6F4), member 8EE005C9-1F88-424E-896D-7600B54A3B05, set byte offset = 120291328.
Jan 29 09:33:59 MacPro kernel[0]: disk3: 0xe0030005 (UNDEFINED).

These failures were unrecoverable (tried using both DU and DW4) and required the RAID volume to be destroyed. Once taking the disks out of the RAID, all drives passed both the Apple Hardware tests, DW tests, and TechTool Pro tests. I then re-created the RAID stripe with the same 4 drives, and restored data to the volume. The RAID volume then performed without error for 34 days, after which time the exact same problem occured.

I've also had two other RAID failures that didn't require the RAID volume to be destroyed; upon re-boot, the journal was able to repair the data. These two failured happend when logging into an Parallels XP VM that was running from the RAID.

The two hard-crash failures that happened during the backup routine (where the RAID had to be destroyed) failed during the backup of the Parallels hdd file (corrupting it). I have posted in the Parallels forums without any helpful responses (and emailed their support team, but we all know how useless that is!). My backup software vendor (Intego) has responded quicky to my support requests, but tell me that it's a hardware problem, and not their software.

Apple doens't provide any technical support for RAID unless you purchase an Advanced Enterprise support plan (or something similarly named).

The S.M.A.R.T. status of these drives are checked daily (DW4), and have never reported an error.

After both hard-failures, all disks have passed all the diagnostics that I have run. What are the chances that this is indeed a hardware failure??? I have been able to recreate the RAID after both failures, and continue to use it without problems for many weeks (actually, exactly 34 days each time) - writing significant amounts of data to the RAID volume (~60% full) without having any kinds of problems.

There seems to something here that is affecting all of us, and it seems that all that we're going to get for support is 'finger-pointing' and "don't use raid" responses.

During one of the DW repair attempts, I saw an error that read something like "invalid sibling link" on the RAID...

Here's hoping that someone can get to the bottom of this - I'm tired of re-creating the RAID and restoring my data.

Mac OS X (10.4.8)

Happy days are here again! 🙂

When you get through this, and have the new drives, I like to zero the entire drive once. I also like to "benchmark" and put it through a rough set of ZoneBench (Intech Speedtools) tests along with QuickBench. There is even an Integrity and Media test that will help insure you don't have I/O errors (but SoftRAID will let you know) as well as it is able to map out bad sectors on the fly, without a reformat, should the need arise. ZoneBench can help find if there is a "slow" or bad area or zone on the drive, which would affect performance and may point to excessive remapped blocks.

Congrats!

You are welcome, of course. Might want to let the folks at SoftRAID know the outcome.

I know someone that is dealing with a similar but more complex RAID 50 problem and "lost" data that might benefit from trying to convert their drives, too.