I once had a regimented backup system. I would weekly clone my main hard disk. It was painful and meant that I could lose a week’s worth of work, but it was a consistent backup. I did it for years. I never had a catastrophic disk failure, but the system would cover me.
When Time Machine finally reached my machine in the form of Leopard, I actually gave up my cloning ways. I made a two disk raid external disk for backup. That terabyte disk has been my sole means of backup ever since. Time Machine may have interrupted me with its inconvenient backup scheduling, but it seemed to be working. I felt secure enough in that system.
That worked right up until I started hearing an odd hissing sound from my backup raid. It was periodic and the Time Machine partition was MIA. I turned the disk off, waited a couple minutes, turned back on and it returned. THe hissing soon resumed, and the disk once again was no longer on my desktop.
That raid has given up the ghost. It makes a great periodic sound, but stores no data. While raids are more prone to single disk failure, it wasn’t expected. I have to wonder if the regular (sometimes near constant) disk access puts undo stress on the hardware for that purpose. Have any of you lost a hard dive prematurely that you use for Time Machine? Is this a side effect of the way this technology works ironically?
My Time Machine drive is still kicking without a problem. I use a Western Digital My Book Home. 320 GB. Before that I was doing SuperDuper! clones every week as well.
Derik, instead of doing a Raid configuration, perhaps you should invest in a Drobo. That way you won't lose everything if one drive dies. I've never used one personally, but so many people rave about them. Maybe worth looking into?
Time Machine puts way less stress on the drive than the system does, so I doubt it.
The symptoms sound more like a bad power supply. Have you tried a different case?
Lastly, you can never have too many backups.
At least one clone, one incremental on site and one incremental off site. That's a good start.
When I first set up a Leopard Server for our church, I used an external HD with the internal drive on the Mac Mini as a RAID. We kept experiencing disk failures. After numerous hours on the phone with tech support, the bottom line is that Apple doesn't recommend setting up a raid between two different HDs. You can partition one HD and make it a RAID no problem, but those pesky cables somehow make things difficult.
But then what happens if the HD fails completely?
My 500 gig MyBook Pro just died on me. It was two years old and since leopard came out it was only used for regular backup.
I'm definitely getting a Drobo after this.
What kind of RAID are you using? RAID 0, which isn't really "R" (redundant), but instead is for speed? Or RAID 1, where the two drives mirror each other? For RAID 1 - which I think should typically be implemented using two identical drives - you lose no information if one drive fails; same with the Drobo, although technically Drobo doesn't literally use "RAID".
Anyway, for the hard drive, Time Machine should be like any other (fairly) disk-intensive application, no?
You make only 1 backup? Are you crazy? Disks fail no matter what you use them for. I always encourage more backups, instead of/in addition to the greater supposed reliability of using a raid configuration in the single backup. Especially given that storage is so cheap now. At least one backup should be offsite from where you use you computer and local backup. I have suffered disk failures on two systems since I started using Time Machine (which I am sure are not due to Time Machine itself), and it is very reassuring to know that you have a second full backup when the system is being restoring from the other one.
I'm actually now on my third western digital hard drive. They seem to last about a year. My first two drivers were Western Digital Pro drives (500GB). The most recent one they sent me was a 500GB My Studio Edition. We'll see if this fairs any better. I'll let everyone know in a year.
Bottom line. Don't backup to raid 0 drives.
I've been using Time Machine on Leopard almost since it was released with a Maxtor 500gb drive and haven't had any problems. Here's hoping that lasts... :)
My 500 Western Digital drive died about two weeks ago which I initially blamed on Time Machine. I've never had a drive die in under a year. I'll have to rethink it, but I'm glad I have an archive procedure in place so that I did't lose much data.
I've been lucky so far with nary a crash... I use a Drobo as my main storage, with it mirrored to a 2TB WD My Book in raid 0.
Western Digital and Maxtor drives fail more frequently than other reliable brands (Hitachi and Seagate come to mind as two decent brands). With HDD prices falling, there is no excuse to buy bottom-of-the-barrel hard drives.
"I have to wonder if the regular (sometimes near constant) disk access puts undo stress on the hardware for that purpose"
As opposed to the "regular (sometimes near constant) disk access " to the hard drive inside your Mac when the machine is awake and being used?
My WD external recently died, after a mere 3 months in action, so that's being replaced under warranty right now. But i don't believe it was in any way linked to Time Machine, whose small incremental back-ups probably make life easier for your external drive.
What with that WD failing, and my macbook's former Seagate internal, i've got the feeling that American hard-drives are made to fail, and next time i'm buying only Japanese. That's why my Macbook now has Hitachi inside.
Raid 1 is technically not one back up it is two ie mirror on two identical drives. However there is a common power supply and case. If one disk dies the other should be OK. If the power suppy dies then whilst there will be no more writing the data on the drives should be Ok for recovery. I am using a WD Book Studio II (note these can do raid 1 as opposed to single raid 0. Cganging drives is a simple slide in and out process. For nuclear power plant type reliability then triple redundancy is the go. So a raid 0 with say a copy to another disk every so often for critical data. A raid 0 is one backup only but again you should have the original data until you get a new disks. longevity of disks is another matter and is dependant upon a great many things (well a few anyway).
Yup, happened here. LaCie 500Gb drive went dead due to stiction after 9 months. That drive got a workout. Probably copied 10-50 gigabytes every 15 minutes all day.
YUP, I had a MyBook 250GB drive. The drive was 13 months old when it made a few clunks and stopped working. The issue I have is that I put my faith in Time Machine. I had files that I knew were on there and I deleted on my iMac. I knew I could just "go back in time" and retrieve the file I wanted.
SILLY me. Guess we need to backup for Time Machine.
Good advice on buying quality hard drives.