Using internal HD for time machne

[hotrod351]

well ive had two seagate goflex external HD fail in the last 4 or 5 years. was thinking a out using a internal HD as time machine. just wonder if ill get more life out of a internal HD over the external HD. or is that the life you can expect to get out of a time machine HD.

 

[sebseb]

well ive had two seagate goflex external HD fail in the last 4 or 5 years. was thinking a out using a internal HD as time machine. just wonder if ill get more life out of a internal HD over the external HD. or is that the life you can expect to get out of a time machine HD.

I've had a WD drive for over 6 years and it still works perfectly.

 

[Weaselboy]

well ive had two seagate goflex external HD fail in the last 4 or 5 years. was thinking a out using a internal HD as time machine. just wonder if ill get more life out of a internal HD over the external HD. or is that the life you can expect to get out of a time machine HD.

I have never read anything that would make me think an internal would last longer than an external, or even anything about Time Machine that would break drives prematurely. I think you just had some bad luck with drives.

I think most would suggest not using an internal drive as backup as any fault with the machine is more likely to wreck the drive than if it is external to the machine.

 

[mac8867]

I am not a fan of Seagate. I like WD and Toshiba drives. This is a historical tendancy of mine simply because Seagate at one time add higher failure rates.

With that said, an internal drive "typically" will outperform an external drive, that is just a matter of the SATA bus being the fastest way to deliver drive space.

 

[paulrbeers]

well ive had two seagate goflex external HD fail in the last 4 or 5 years. was thinking a out using a internal HD as time machine. just wonder if ill get more life out of a internal HD over the external HD. or is that the life you can expect to get out of a time machine HD.

My honest opinion is that cheap external hard drives (by Seagate or WD) have poor cooling. They are generally plastic casings with no active cooling. Last time I was tracking it, WD and Seagate would give 3/5 year warranties on internal drives but only 1 year on external. While that doesn't necessarily mean they expect externals to fail more often, it definitely makes me question the build quality. If it is a mission critical drive, I would recommend buying a "bare" drive (internal) and a good external case that is either Metal (Aluminum preferred) or with active cooling (or both!).

I actually buy cheap externals, rip the drive out of them and put them into 4 bay enclosures (it is usually cheaper to buy externals than bare... go figure?). Downside is that I obviously void my warranties with this, but I couldn't tell you the last time I had a failure in less than a year using this process (which is all the warranty would be on those external drives).

Food for thought....

Edit: I've had pretty good luck with these: http://www.amazon.com/Mediasonic-Pr...d=1417452738&sr=8-1&keywords=mediasonic+4+bay

I did have one bad one, but Amazon took it back and sent me a new one with no issues and it was clearly a dud from the beginning. On the 3 I have kept, they have all been running 24/7 for a couple of years with no issues.

 

[chabig]

I think most would suggest not using an internal drive as backup as any fault with the machine is more likely to wreck the drive than if it is external to the machine.

Computers sometimes fail, but they seldom explode or catch fire. I think an internal drive is perfectly safe and I'd have no qualms about using one for backup.

Something the OP should remember, is that one backup is not sufficient. He should have multiple backups and keep one offsite, rotating them periodically.

 

[Weaselboy]

Computers sometimes fail, but they seldom explode or catch fire. I think an internal drive is perfectly safe and I'd have no qualms about using one for backup.

How in the world did you get anything close to that from what I said?

 

[Cubytus]

I am not a fan of Seagate. I like WD and Toshiba drives. This is a historical tendancy of mine simply because Seagate at one time add higher failure rates.

Just a matter of luck. I have Seagate and WD drives, internal WD replaced twice in three years, Seagate twice in two years, Hitachi once in two years. Current mechanical HDDs are basically junk.

On the other hand and contrary to my expectations and previous experience, Japanese-branded HDDs seem to last longer. This wasn't the case a few years ago.

With that said, an internal drive "typically" will outperform an external drive, that is just a matter of the SATA bus being the fastest way to deliver drive space.

True, but probably not a critical point as far as backups are concerned.

My honest opinion is that cheap external hard drives (by Seagate or WD) have poor cooling. They are generally plastic casings with no active cooling. Last time I was tracking it, WD and Seagate would give 3/5 year warranties on internal drives but only 1 year on external. While that doesn't necessarily mean they expect externals to fail more often, it definitely makes me question the build quality. If it is a mission critical drive, I would recommend buying a "bare" drive (internal) and a good external case that is either Metal (Aluminum preferred) or with active cooling (or both!).

I have an external, metal case holding a Seagate. Powered about 6 hours a week. Still failed just over 2 years of age.

If you can afford it, take a RAID1 enclosure. Sure drives will still be crappy, but I think their risk of failing
at exactly the same time is rather slim.

 

[chabig]

How in the world did you get anything close to that from what I said?

You said that a machine fault would be more likely to "wreck" an internal drive than an external one. That implies physical risk.

 

[Weaselboy]

You said that a machine fault would be more likely to "wreck" an internal drive than an external one. That implies physical risk.

I was referring to electrical problems like power supply problems or even a power surge. But I suspect you knew that.

 

[chabig]

I was referring to electrical problems like power supply problems or even a power surge. But I suspect you knew that.

I will agree with you on those concerns. I wouldn't have thought of them myself. As a devil's advocate, I'd probably put more faith in the quality of the Mac Pro's power supply than those that power many external drives.

 

[Cubytus]

I was referring to electrical problems like power supply problems or even a power surge. But I suspect you knew that.

Physical problems are not excluded either. Laptops take a beating being hauled from one place to the other, and even with precautions, they wear out quicker than desktops.

 

[ApfelKuchen]

A key point to a backup is to be external to the machine. If it's not going to be off-site, then, imo, it should be quickly and easily separated from the machine to speed disaster recovery.

With an external drive, it's plug-n-play into whatever machine you're migrating. With an internal drive, besides the extra work of removing it (presuming it's a machine failure, rather than a boot drive failure), will you have a spare external housing/adapter handy for migration to another machine, or must the next machine have a spare internal drive bay (not exactly something to expect from a new Apple computer)?

My feeling is, so what if external drives may not last as long? It's a fairly small cost of doing business compared to the value of having that backup, if needed.

Does the external housing contribute to shorter life span? The drives are the same as internals (or ought to be), so we're really talking about whether the external housing provides less cooling, or the external's power supply is dirtier than the internal supply. What other variables would have an impact on HDD longevity? Backup drives generally don't work very hard, so not much heat generated there. Wall wart power supplies don't contribute to internal heating at all... My external backup feels cooler than my very cool iMac.

 

[mac8867]

A key point to a backup is to be external to the machine.

This is very true, no argument. The OP did say however that this was for Time Machine. I wouldn't (and do not) consider Time Machine to be a backup. Meaning, I do not use it as the basis of recovery, it is simply the interim snapshots between full backups. I DO recommend backing up your Time Machine/Capsule drive though!

 

[hotrod351]

i have two 750gb internal HD's. i cloned one to the other and keep it as a backup in case the one im working off of fails. guess i could just use CCC every once in a while to clone it for backup.

 

[Cubytus]

i have two 750gb internal HD's. i cloned one to the other and keep it as a backup in case the one im working off of fails. guess i could just use CCC every once in a while to clone it for backup.

That's a good first start, but you probably don't need versioning, then?

This is very true, no argument. The OP did say however that this was for Time Machine. I wouldn't (and do not) consider Time Machine to be a backup. Meaning, I do not use it as the basis of recovery, it is simply the interim snapshots between full backups. I DO recommend backing up your Time Machine/Capsule drive though!

Ok, and if TM is not a backup, what is it then?

And what is your definition of backup?

 

[ApfelKuchen]

This is very true, no argument. The OP did say however that this was for Time Machine. I wouldn't (and do not) consider Time Machine to be a backup. Meaning, I do not use it as the basis of recovery, it is simply the interim snapshots between full backups. I DO recommend backing up your Time Machine/Capsule drive though!

There's a bit of room for variation in backup strategies. It seems likely that for the vast majority of its users, Time Machine is their only backup. A good bit of the point on Apple's end is to provide an "it just works" solution for those who weren't backing up at all.

For those who have always done backups (dare I say, industrial-strength backups), sure, Time Machine can be a nice supplement. It's just not the only way Time Machine can be (or is) used.

 

[Cubytus]

There's a bit of room for variation in backup strategies. It seems likely that for the vast majority of its users, Time Machine is their only backup. A good bit of the point on Apple's end is to provide an "it just works" solution for those who weren't backing up at all.

For those who have always done backups (dare I say, industrial-strength backups), sure, Time Machine can be a nice supplement. It's just not the only way Time Machine can be (or is) used.

Again, would you elaborate on what you call "industrial-strength" backups for the average power user? Using a RAID1 as a destination rather than a single drive? Geographical separation? Bootable backup?

 

[ApfelKuchen]

Again, would you elaborate on what you call "industrial-strength" backups for the average power user? Using a RAID1 as a destination rather than a single drive? Geographical separation? Bootable backup?

I'd say that any strategy that has the goal of near-zero risk of data loss AND minimal disruption fits the bill. It involves at least two independent, full-backup methods, at least one of those off-site. Risk and the manner in which risk is addressed is situational, so it's hard to itemize method. There's interplay with the importance of continuous uptime/tolerable amounts of downtime, economics...

To my line of thinking, RAID is not backup, it's fault tolerance in situations where you want to maximize uptime. "Backup" is what you have in case the RAID is destroyed in a fire or flood.

I'm not sure I'm a power user (definitely not as much as some, definitely more so than others). I consider my backup approach to be garden variety, rather than industrial strength - a local, external Time Machine backup on a consumer-grade USB HDD, plus iCloud and Dropbox for critical data. I know the risks, I know the costs, I know I'll be able to cope with the consequences. I take a far more meaningful risk every time I jump into my car.

 

[26139]

HUh?

Okay, so I assume you have two internal HDs installed, yes?

And you want to use one for backup.

Okay, so that covers HD failure, but what if the computer is damaged? Stolen?

This just seems an odd choice to me.

 

[Cubytus]

I'd say that any strategy that has the goal of near-zero risk of data loss AND minimal disruption fits the bill. It involves at least two independent, full-backup methods, at least one of those off-site. Risk and the manner in which risk is addressed is situational, so it's hard to itemize method. There's interplay with the importance of continuous uptime/tolerable amounts of downtime, economics...

To my line of thinking, RAID is not backup, it's fault tolerance in situations where you want to maximize uptime. "Backup" is what you have in case the RAID is destroyed in a fire or flood.

I'm not sure I'm a power user (definitely not as much as some, definitely more so than others). I consider my backup approach to be garden variety, rather than industrial strength - a local, external Time Machine backup on a consumer-grade USB HDD, plus iCloud and Dropbox for critical data. I know the risks, I know the costs, I know I'll be able to cope with the consequences. I take a far more meaningful risk every time I jump into my car.

Wrong: RAID1 on the backup drive makes a total of 3 drives. But it does cover near-zero data loss risks if faulty disk is replaced as soon as failed, doesn't it?

Fire & Flood = geographical separation, then? Your strategy doesn't seem to include full backups offsite, and that was also my main issue: not being able to have a full offsite backup. Cloud providers just don't have the necessary space, nor privacy. I uninstalled Dropbox when I learnt about hubiC, that now offers backups (automated or manual) and have very large spaces available for so few bucks. In which case the upload bandwidth becomes a limiting factor for the first backup. They also have a plugin available if your local backup happens to be a Synology enclosure.

I calculated that my strategy allows me to be running again in about 30 min for critical documents, overnight for full restore, with a maximum loss of about 12 hours of work. This was before SpiderOak's utter failure to repair my space, and before the single backup drive unexpectedly failed despite really low usage.

The problem is that cheap backup is affordable, but prices climb very quickly once you factor in options such as NAS, external providers, and other, distant NAS.

 

[ApfelKuchen]

Wrong: RAID1 on the backup drive makes a total of 3 drives. But it does cover near-zero data loss risks if faulty disk is replaced as soon as failed, doesn't it?

I was "wrong" to assume that you were using RAID1 for primary storage. You earlier asked, "...Using a RAID1 as a destination rather than a single drive?" If you'd said 'Using a RAID1 as a backup,' your meaning would have been more clear. Still, that begs the question of why you'd have a single drive as primary storage, backed-up by a RAID1? You've created a more reliable backup, but have done nothing for the reliability of primary storage. If the 'best backup' can be defined as, "The backup you never have to use," then primary RAID1, single-disk backup makes more sense (to me) as an allocation of three HDDs.

Fire & Flood = geographical separation, then? Your strategy doesn't seem to include full backups offsite, and that was also my main issue: not being able to have a full offsite backup.

I guess you mean my personal backup strategy, which I already admitted does not qualify as industrial strength. I was quite clear that my definition of industrial strength does include full off-site backup.

Cloud providers just don't have the necessary space, nor privacy.

Off-site these days implies 'cloud,' but it's not the only method by any means. When I was doing corporate IT back in the 90s, it meant sending a streaming tape of our RAID5 server to an offsite repository on a regular schedule. There are those who take a portable drive home with them at the end of the work day, others deposit those portable drives in bank vaults on a regular schedule (since theft is also a consideration, encryption of those backups seems a good idea). Real-time off-site backup has advantages over a periodic snapshot of that sort, but even then, it can be a matter of using a VPN to send the data to another site under your control, rather than put it in the hands of strangers.

I uninstalled Dropbox when I learnt about hubiC, that now offers backups (automated or manual) and have very large spaces available for so few bucks. In which case the upload bandwidth becomes a limiting factor for the first backup. They also have a plugin available if your local backup happens to be a Synology enclosure.

I calculated that my strategy allows me to be running again in about 30 min for critical documents, overnight for full restore, with a maximum loss of about 12 hours of work. This was before SpiderOak's utter failure to repair my space, and before the single backup drive unexpectedly failed despite really low usage.

The problem is that cheap backup is affordable, but prices climb very quickly once you factor in options such as NAS, external providers, and other, distant NAS.

I'm not going to say "industrial strength" means "cost is no object," but an industrial enterprise probably does have more money at stake, so their cost-benefit analysis can allow for solutions that are prohibitively expensive for the single user/very small business.

In the end, it's not whether a backup is "industrial strength" by my casual definition or any other - it's whether the strategy selected is appropriate to a user's needs, based on a well-reasoned assessment of risk and cost.

If there's an ISO standard that includes backup strategies, that would be a better measure of "industrial strength" than anything we can debate here.

 

[Cubytus]

I was "wrong" to assume that you were using RAID1 for primary storage. You earlier asked, "...Using a RAID1 as a destination rather than a single drive?" If you'd said 'Using a RAID1 as a backup,' your meaning would have been more clear.

That is exactly the meaning of my sentence. Destination is of course the backup drive, not the primary drive.

Still, that begs the question of why you'd have a single drive as primary storage, backed-up by a RAID1? You've created a more reliable backup, but have done nothing for the reliability of primary storage. If the 'best backup' can be defined as, "The backup you never have to use," then primary RAID1, single-disk backup makes more sense (to me) as an allocation of three HDDs.

That would be in fact quite good, but just impossible in a laptop for space reasons.

True a good backup is one that never needs to be used, and such was the case of my Time Machine drive when I discovered it wouldn't mount, wouldn't repair for unknown reasons. Then I had the time to do a CCC on an external drive, but even this is clearly annoying as it means additional maintenance.

Off-site these days implies 'cloud,' but it's not the only method by any means. When I was doing corporate IT back in the 90s, it meant sending a streaming tape of our RAID5 server to an offsite repository on a regular schedule. There are those who take a portable drive home with them at the end of the work day, others deposit those portable drives in bank vaults on a regular schedule (since theft is also a consideration, encryption of those backups seems a good idea). Real-time off-site backup has advantages over a periodic snapshot of that sort, but even then, it can be a matter of using a VPN to send the data to another site under your control, rather than put it in the hands of strangers.

So for the common people, that would take the form of a WAN-accessible NAS located at a friend's house? This would cover the "real-time" replication of data, but unfortunately requires some hefty Internet connection.

And encryption of backups still doesn't seem to be commonplace: SpiderOak has an awful performance, Time Machine only supports it in Time Capsule, and other vanilla cloud providers don't encrypt. Do consumer-level NAS do it?

 

[Weaselboy]

And encryption of backups still doesn't seem to be commonplace: SpiderOak has an awful performance, Time Machine only supports it in Time Capsule, and other vanilla cloud providers don't encrypt. Do consumer-level NAS do it?

I have been using the app Arq to backup to Amazon's S3 servers. Cost is three cents per month per GB. Arq encrypts the backup before it leaves my Mac. I have been pretty pleased with the setup so far.

In my tests I have been able to saturate my 30/5 connection in both directions.

 

[Cubytus]

I have been using the app Arq to backup to Amazon's S3 servers. Cost is three cents per month per GB. Arq encrypts the backup before it leaves my Mac. I have been pretty pleased with the setup so far.

In my tests I have been able to saturate my 30/5 connection in both directions.

Looks interesting, but has it been audited? This is important for a security software.

Never mind the fact that it wouldn't work on my daily driver Snow Leopard.

Amazon S3 seems ok for its granularity, but hubiC is still cheaper