Apple RAID Frustrations and Request for Advice

[Ledgem]

Greetz all,

I have two 1.5 TB drives in a RAID 1 (mirror) array, set up through the Disk Utility. I don't care about performance. These drives store my photos and videos that are still being processed; I keep copies of the final versions on the RAID and on other external drives. The point of the RAID is to guard against losing anything "in progress" before I get around to copying it to other drives.

At this point, less than six months in with the array (and fairly light usage), I've had to build the array three times. I unmount the drives properly and have no idea why the RAID seems to degrade so frequently. Oddly, one time Mac OS X reported that the array was degraded, but after dismounting the drives, shutting them off, and then turning them back on in reverse order, it reported that the RAID was healthy. It's strange, it doesn't lend much confidence to the robustness of the RAID (in my opinion), and I'm getting sick of losing time on the rebuilds.

I'm curious: does anyone have any theories as to why the RAID is degrading? I'm also curious if anyone has any better suggestions. I could run the drives separately and use some sort of synchronization software, I suppose, but I'm unfamiliar with any such software, nor do I know what the limitations would be. I presume that they don't synchronize as quickly as you would get with a RAID, and perhaps they would take more resources...?

Any thoughts, suggestions, commiserations, etc. are welcome.

 

[Mattie Num Nums]

Greetz all,

I have two 1.5 TB drives in a RAID 1 (mirror) array, set up through the Disk Utility. I don't care about performance. These drives store my photos and videos that are still being processed; I keep copies of the final versions on the RAID and on other external drives. The point of the RAID is to guard against losing anything "in progress" before I get around to copying it to other drives.

At this point, less than six months in with the array (and fairly light usage), I've had to build the array three times. I unmount the drives properly and have no idea why the RAID seems to degrade so frequently. Oddly, one time Mac OS X reported that the array was degraded, but after dismounting the drives, shutting them off, and then turning them back on in reverse order, it reported that the RAID was healthy. It's strange, it doesn't lend much confidence to the robustness of the RAID (in my opinion), and I'm getting sick of losing time on the rebuilds.

I'm curious: does anyone have any theories as to why the RAID is degrading? I'm also curious if anyone has any better suggestions. I could run the drives separately and use some sort of synchronization software, I suppose, but I'm unfamiliar with any such software, nor do I know what the limitations would be. I presume that they don't synchronize as quickly as you would get with a RAID, and perhaps they would take more resources...?

Any thoughts, suggestions, commiserations, etc. are welcome.

Have you tried changing the block size? I had a RAID that was heavy on multimedia and it degraded frequently until I increased the the block size.

 

[Ledgem]

Have you tried changing the block size? I had a RAID that was heavy on multimedia and it degraded frequently until I increased the the block size.

I'll give it a try. What block size would you suggest?

 

[Mattie Num Nums]

Depends on the data. If its video you want the largest chunk possible. If its data like docs and stuff it doesn't matter much. I usually set it to 128K for mixed and 256k for mixed with multimedia or large files.

 

[Ledgem]

Finally cleared up the space to take the data off of the RAID in order to break it and remake it. Ran into an interesting issue: making a RAID 1 with a 256K block size causes kernel panics. It causes a kernel panic when the RAID is being created (at the second to last step), and if you leave the drives on and connected, it causes kernel panics at the start-up screen (Apple logo), too. If you connect both drives, the system throws a panic as it attempts to bring the RAID online. The work-around is to turn on only one drive at a time, and erase it. I tried connecting the drives in various ways, but it didn't help - it seems to be an OSX issue. I don't know how widespread it is, but I was able to find other posts from people with similar issues as far back as 10.5. That's disappointing.

Instead, I've settled on a RAID 1 with a 128K block size. If I still have frequent degrading issues, I'll report back here. Thanks again for the advice, Mattie.

 

[fromeout11]

Same issue

After quite a bit of frustration, I discovered that I have the same issue... 256k block size RAID mirroring causing consistant kernel panics on 10.8.3. Apple needs to squash this bug in Disk Utility.