S.M.A.R.T status alerts from disks in RAID?

[cb911]

I was just seeing if I could see the S.M.A.R.T status of the disks that I'm using in RAID. But I can't. It looks like the only way is through the RAID Controller? So I've setup email notification of alerts & errors at least, but is there anyother way I can monitor the status without having to log-in to the controller?

Also - is there no S.M.A.R.T status Dashboard widget? The only related app I found was SMART Reporter for OS X.

Just hoping that someone has more experience with monitoring S.M.A.R.T status on a RAID more than I do. Any help would be great.

 

[Eniregnat]

Humm, I hope somebody can come up with a better answer, although, it is unlikely that any more than 1 drive will fail with in a RAID, and while the SMART status of any one drive can predict failure, and help you manage your RAID, it might be over kill. (Please note this is ironic, as I administrate a RAID 5 with hot spare, that is also mirrored on an identical RAID 5. Nearly zero down time, and a lot of redundancy.)

[Blather below is not meant to be insulting. I am just thinking as I type and cut and paste from previous posts.]

S.M.A.R.T = Self-Monitoring, Analysis, and Reporting Technology. There is not SMART status for a RAID, as SMART referers too and is only about single drives. It would be a great tool for RAID management, but perhaps it is a little over kill. Apple Remote Desktop should allow you to check individual disks with in a RAID, but this isn't what your looking for.

OS X Server, offers this level of micro management through remote diagnostics. There are a lot of RAID and management features that are not easily accessed in regular old OS X.n, it's how they sell the Server version.

I don't know how to access in OS X. Perhaps a script that quires the DiskUtility and reports if the status is not 'x'. I don't have my Mac in front of me, so I can't see what features can be scripted to work with the DiskUtility. Of course this is a rather involved script. Each and every disk in the RAID would have to be queried through a script, and that scrip would have to be automated to run at regular intervals. I don’t know how I would test the script?

[Brain Burp Below, for those who do not know what SMART status is.]
What SMART will report is a missed sector, or a discovered bad sector that can not be recovered, is not good. The information is logged and the error mapped, diminishing usable drive capacity. SMART will report that the drive is failing. More than a few of these and the drive is reported as FAILED- but usable. SMARTs purpose is to help predict dive life and to help users predict drive imminent drive failure.

Most drive platters have imperfections that are mapped when the drive is at the factory. The information is stored on the drive so that data is not stored in those locations. This 'error' map is not lost during a format. Manufactures grade platters, the platters with more imperfections platters go into less expensive drive assemblies. SMART allows the ‘error’ map to more dynamic.

Links SmartTools GPL
Sorry- random brain burp.

 

[robbieduncan]

I think you can only query SMART status over a local IDE (or SATA) bus.

 

[cb911]

Thanks for the replies. Just a few things though - I'm wondering how Apple Remote Desktop would let you check individual disks used in a RAID? Since the disk appears as one to the OS?

I'm thinking that the only way to make this kind of widget would be to have the widget login to the Controller Card then poll the S.M.A.R.T status this way. Sounds like a complicated bit of coding, too much for me.

But as you mentioned, this may be overkill on a RAID. I guess the only thing that has me worried is that I've used all the same model/batch Maxtor drives. Some people love 'em, some people hate 'em. I'm just worried about the chances of more than 1 going at a time. But, RAID isn't a solution for backup, so I guess I'll just try to keep everything backed up and worry about the drive failure when it happens.

Thanks for the link as well, there was a link on that page to smartmontools for SCSI disks, which is what OS X recognizes the RAID as, so there may be some help there.

 

[yippy]

Yea, only over a local bus. You can't even get to it on a normal firewire or USB hard drive. That you can get to it at all through the raid controller I would say is pretty good.

Sorry, not much help but that is what I know about smart (I don't know anything about using a RAID array other than the basic principle).

 

[cb911]

Sorry, I may not have explained with enough details...

I can view the S.M.A.R.T status for each device if I log into the Controller Card - here you can view each device (individual hard drive), as this is also where you create and manage the RAID arrray. So when I say through the RAID Controller, I mean that you've got to login to the web management utility to view the status. All I'm looking for is an easier way than that.

The RAID array does appear to OS X as a SCSI device.

Hrmm... hope that explains it better.

 

[Eniregnat]

Thanks for the replies. Just a few things though - I'm wondering how Apple Remote Desktop would let you check individual disks used in a RAID? Since the disk appears as one to the OS?

The only way ARD would let you view individual drives is if there was a utility program (yours is web based) for the SCSI controller. If you can directly address your machine from over the net, then ARD would be the way to go. It's over kill, and as far as I know it, you still have to use the web management utility. I didn't understand that. Again, I think the tools are built into the OSX Server, not the regular OS X, but I could be wrong. Others are correct, in that indirectly accessed drives, those moderated by intermediary system like FW or USB, will not display the SMART status. A caveat to this is that the SMART data is still being collected and stored on the drives tiny flash memory or on the disk directly. So pop it out of the external enclosure and directly attach it to an interface card, and you will see good SMART data. (A pointless amount of work.)

If that utility can be scripted, then you could do the work. For now, if you just check once a week, you should be fine.

About my RAID expirence.

The NAS RAIDS I work with are composed of Maxtor drives, and I have had failures, but only single drives at a time. They are admind via a web interface, and the OS is JAVA based (A pain to upgrade, but I still love Sun).

Basically the reason they failed was that the NAS's were purchased as refurbished, and they never really fixed the original problem. Rebuilding my RAIDS should a disk fail, can take up to 22hrs, even if off line. That's why I have a spare mirror of the primary RAID. It's really way too redundant, but I work with media that can not be easily recreated. Time is money, so dual redundant raids is very cost effective in the long run. I will note, that when my primary RAID drive has a failure, I back it off the network, do a final sync with the backup mirror, reverse the addressing, and reverse their rolls until I get a replacement drive and controller. We pay for 24hr technical support, but after years of sending detailed logs back and forth, they assume that we know what a failure is, and overnight us a replacement drive for the failing RAID. Rebuilding a RAID can be painful, and I hate running anything in a degraded state.

 

[cb911]

damn you must have alot of drives/capacity in that RAID.

I guess I will just check the status every week or so. I am paranoid about losing my data, but I guess that alone won't stop anything from happening.