Tech myth debunk thread

[Lady Fitzgerald]

24 minutes ago, mr moose said:

Any issue with files, software or user error and your data is not backed up.

 

8 minutes ago, Drak3 said:

That's true of any backup system. NAS and versioned libraries are not failsafe either.

No, that is not true. If you have a true backup of your data that is kept disconnected from the computer, powered down, and stored out of sight of the computer, you most definitely can recover data lost due to system hiccups or user error (and I have done so in the past). If using the correct software that's been configured correctly to update your backups, you can even keep multiple versions of changed data

 

The purpose of redundancy is to allow a computer to keep chugging along without data loss if a drive should go belly up, which is a perfectly valid use for it. However, since it refreshes every time data is changed, data that becomes corrupted or deleted will be lost. Only an up to date backup that is kept disconnected from the computer, powered down, and stored away from the computer is capable of restoring lost data, no matter how it gets lost.

[mr moose]

18 minutes ago, Drak3 said:

That would be a type of backup. You have dedicated a set of drives to being local and live backups of each other.

Except that the data is only duplicated not backed up.  There is a big difference between backed up and duplicated.  Files are not backed up if they cannot be recovered in the event of a software failure.

 

18 minutes ago, Drak3 said:

That's true of any backup system. NAS and versioned libraries are not failsafe either.

Point?

18 minutes ago, Drak3 said:

https://www.merriam-webster.com/dictionary/backup

These are the two relevant definitions. A drive failure results in the practical data loss of that drive.

RAID 1 is a mirror arrangement where the purpose is that data is not lost if a drive fails. By both definitions, RAID 1 is a form of backup. A scenario specific backup.

But the data is lost in anything except a drive failure.  The definition of a backup is literally:

 

Quote

to make a copy of (a computer file or data) to protect against accidental loss or corruption

 

 

No raid will protect you against accidental loss or corruption and it's questionable if it will even save you from corrupted files.

 

Quote

 

18 minutes ago, Drak3 said:

https://www.merriam-webster.com/dictionary/redundant

And here's the most applicable definition of redundant.

Any backup exists as duplicate data to get a system running again. Different types of backups do so in different capacities. RAID 1 enables a system to be restored with little downtime in the event of a drive failure. That's what it's good for, that's what it should be used for. It's redundant, as every backup system is, by definition.

 

 

Semantics again, you are trying to stretch the definition of backup to include something that does not fits it's description nor usage.     Why in the face of god knows how many links to relevant articles, nearly every definition presented (including your own) and half a dozen people telling you that a raid is not a backup,  do you insist it is?

 

 

 

 

 

[Lady Fitzgerald]

32 minutes ago, amdorintel said:

...5. Portability:  Tape is still the only medium on which you can reliably store data and move it offsite for disaster recovery purposes...

Excellent article. However, the above is not quite true. Solid State storage is also very portable. However, it does not have the longevity of tape (as pointed out later). Still, it is useful for short term storage and is far less expensive to use on a smaller scale.

 

24 minutes ago, Kisai said:

It's a pity that there's no real alternative to tape, though flash media might actually get there for short-term backups (eg 1-5 years.)

Five years for SSDs is a bit of a stretch (flash media, such as USB thumb drives or camera cards, usually have even less data retention durability). Even one year might be a bit optimistic, depending on the age of the drive, how full it is, and the number of bits store per cell.

 

I refresh all my data SSD backups no less than once a month, even if no data has been added or changed. While that's rather paranoid, I swap out my offsite backup drives with my onsite backup drives also no less than once a month to keep the offsite backup drives as up to date as possible so it's a convenient time to refresh them when updating the ones that have had changes. Refreshing is as simple as connecting the backup drive to the computer and reading it (I use my data backup program, FreeFileSync, to compare the computer data drive to its backup drive to ensure the drive gets read; it also can catch any "bit rot", which hasn't been a problem for me yet). I can easily do all ten 4TB backup SSDs while watching the evening news.

[Drak3]

1 minute ago, Lady Fitzgerald said:

The purpose of redundancy is to allow a computer to keep chugging along without data loss if a drive should go belly up,

So, fitting the definition of backup.

 

1 minute ago, Lady Fitzgerald said:

No, that is not true

Yes, it is. No backup is truly safe from damage in some form.

 

I thought you knew that, considering how concerned you are with house fires, and how you spoke down upon another member when they bought a firesafe to store an external drive.

1 minute ago, mr moose said:

There is a big difference between backed up and duplicated.

No, there isn't. Some backup systems are just multiple duplicates of of the drive, sometimes not even on a separate disk (meaning partitions, which is fine for basic safe guarding against corruption).

 

1 minute ago, mr moose said:

But the data is lost in anything except a drive failure.  The definition of a backup is literally:

 

Quote

to make a copy of (a computer file or data) to protect against accidental loss or corruption

 

 

No raid will protect you against accidental loss or corruption and it's questionable if it will even save you from corrupted files.

 

RAID protects against loss from drive failures, meaning that still fits that definition. It is literally a copy of a drive to protect against the failure of said drive.

 

2 minutes ago, mr moose said:

ou are trying to stretch the definition of backup to include something that does not fits it's description nor usage.

No, I'm not. I'm taking a word, and using it for its actual definitions.

3 minutes ago, mr moose said:

nearly every definition presented (including your own)

The ACTUAL definitions presented don't contradict what I said.

 

4 minutes ago, mr moose said:

half a dozen people telling you that a raid is not a backup

Because those people haven't given a counterargument to the actual definitions. You evade or dismiss outright the definition.

 

And this is why I separate Mira as being one of the smartest members of the forum. From the get go, they acknowledged that they had their own concept of what a backup was, and didn't treat it as absolute. After some discussion, they acknowledged that there is a distinction to be made between types of backups, that you can have backups for hardware and backups for data, and that RAID doesn't really handle the task of data backup. Mira 'listened' to those with different viewpoints, considered them, and adjusted their own accordingly.

 

Whereas multiple people have been shown the actual definition, how this type of backup is used, and ignore it. As far as I'm concerned, you'd rather try to hide behind """expert""" opinion pieces and trying to frame the discussion as "semantics," than do any critical thinking.

[Kisai]

3 minutes ago, Lady Fitzgerald said:

 

Five years for SSDs is a bit of a stretch (flash media, such as USB thumb drives or camera cards, usually have even less data retention durability). Even one year might be a bit optimistic, depending on the age of the drive, how full it is, and the number of bits store per cell.

 

I was thinking incremental backups, eg one monthly backup with daily/weekly increments backup's to SSD and then when it gets to the end of the month move do a new monthly backup, and offsite the previous months backups. Then keep whatever level of redundancy you need after that. So if you need 60 SSD's to do two months, you can either reuse the drives a few times, or recycle/wipe them like with tapes. A "smart" e-waste friendly variation would be to only reuse the drives until they hit 50% of their expected TBW and then replace them. 

 

Because the office basically needs to be recoverable and not lose days of work, however way they do backups (I literately have no idea how it works at present, as they replaced two systems connected by fiber channel to the Quantum Tape drive. I know "cloud" is involved at some point) wherever these tapes are stored would have to be retrievable, locally, so I have no idea what their disaster recovery plan is. I've never seen anyone try to restore a backup.

 

At any rate Holographic storage when? Who's sitting on it?

[Lady Fitzgerald]

5 minutes ago, Drak3 said:

So, fitting the definition of backup.

 

Yes, it is. No backup is truly safe from damage in some form.

 

I thought you knew that, considering how concerned you are with house fires, and how you spoke down upon another member when they bought a firesafe to store an external drive.

No, there isn't. Some backup systems are just multiple duplicates of of the drive, sometimes not even on a separate disk (meaning partitions, which is fine for basic safe guarding against corruption).

 

RAID protects against loss from drive failures, meaning that still fits that definition. It is literally a copy of a drive to protect against the failure of said drive.

 

No, I'm not. I'm taking a word, and using it for its actual definitions.

The ACTUAL definitions presented don't contradict what I said.

 

Because those people haven't given a counterargument to the actual definitions. You evade or dismiss outright the definition.

 

And this is why I separate Mira as being one of the smartest members of the forum. From the get go, they acknowledged that they had their own concept of what a backup was, and didn't treat it as absolute. After some discussion, they acknowledged that there is a distinction to be made between types of backups, that you can have backups for hardware and backups for data, and that RAID doesn't really handle the task of data backup. Mira 'listened' to those with different viewpoints, considered them, and adjusted their own accordingly.

 

Whereas multiple people have been shown the actual definition, how this type of backup is used, and ignore it. As far as I'm concerned, you'd rather try to hide behind """expert""" opinion pieces and trying to frame the discussion as "semantics," than do any critical thinking.

At this point, I've decided you must be trolling (posting to deliberately stir up discord). And stop twisting my words around and taking them out of context.

[mr moose]

6 minutes ago, Drak3 said:

 

RAID protects against loss from drive failures, meaning that still fits that definition. It is literally a copy of a drive to protect against the failure of said drive.

 

 

No, I'm not. I'm taking a word, and using it for its actual definitions.

The ACTUAL definitions presented don't contradict what I said.

 

Because those people haven't given a counterargument to the actual definitions. You evade or dismiss outright the definition.

 

And this is why I separate Mira as being one of the smartest members of the forum. From the get go, they acknowledged that they had their own concept of what a backup was, and didn't treat it as absolute. After some discussion, they acknowledged that there is a distinction to be made between types of backups, that you can have backups for hardware and backups for data, and that RAID doesn't really handle the task of data backup. Mira 'listened' to those with different viewpoints, considered them, and adjusted their own accordingly.

 

Whereas multiple people have been shown the actual definition, how this type of backup is used, and ignore it. As far as I'm concerned, you'd rather try to hide behind """expert""" opinion pieces and trying to frame the discussion as "semantics," than do any critical thinking.

 

yep, clearly everyone else is wrong.

 

 

 

[Drak3]

Just now, mr moose said:

 

yep, clearly everyone else is wrong.

 

 

 

Argumentum Ad Populum.

 

You haven't presented a counterpoint as to why RAID 1 doesn't fit the definition of backup. You just summarized what I said: RAID 1 only covers data loss that would result from a disk failure.

[mr moose]

20 minutes ago, Drak3 said:

Argumentum Ad Populum.

 

You haven't presented a counterpoint as to why RAID 1 doesn't fit the definition of backup. You just summarized what I said: RAID 1 only covers data loss that would result from a disk failure.

 

Yes I have, and so have many others:

 

got back to here and start again:

 

6 hours ago, wkdpaul said:

No , I'm not.

 

RAID isn't a backup, like I said, it's redundancy.

 

RAID won't protect you form viruses, malwares or cryptolockers, etc.

RAID won't protect you form mistakes (people DO delete files by accident).

RAID won't protect you from data corruption.

 

All of the above will replicate the problem to all drives in the RAID.

 

 

I have a backup solution, the first step is to copy the files on a RAID array (the array is part of my backup system, but the drives where the data is mirrored is NOT a backup, it's a Fail-Safe in case of drive failure), the second step is to do an incrememtal cloud backup, last step is to copy only specific "life-or-death" types of files to 2 different off-site and cloud storage.

 

[leadeater]

6 hours ago, amdorintel said:

tape is like so old, so so old

Tape technology is old, tape drives and tape storage devices under go generational improvements. Saying tape is old as a negative does not apply because you can say that about disks, memory and flash storage. All these technologies are extremely old yet there are new advancements with them every year.

 

We used to have thousands of LTO-5 tapes, now we have hundreds of LTO-7 tapes, which have already been superseded by LTO-8 (we will skip this generation and go LTO-9/LTO-10). A current generation LTO-8 tape can hold up to 30TB of data with a data rate between 360-900 MB/s, and if you're like us you'll have many of these tape drives so can be coping upwards of 10 GB/s, that means you need extremely fast storage arrays to actually be able to feed the data fast enough to the drives or you'll cause tape scrubbing.

 

Tape is still by far the cheapest backup storage medium as all it requires is a single drive and as many tapes as you need and time to swap tapes, or go higher end and use auto-loading tape libraries. Only thing cheaper is abusing free cloud storage, but those are extremely slow.

[leadeater]

5 hours ago, Kisai said:

Also, stupidly expensive IMO. It's a pity that there's no real alternative to tape

Well who on earth buys a tape drive if your backup requirements don't even fill a single tape, going back years that used to be the only option but now for those use cases we have cloud storage. You can get an LTO-7 drive for about $2k and each tape after that is between $40-$60 so it's not expensive but rather very cheap so long as you have the data footprint to meet that minimum entry point. I know it's a technicality here because barrier of entry might seem high, and it is for home usage, but even small businesses today have a fair decent amount of data so tape is a front runner for cost for almost everyone. There's more important factors that would push these people away from tape though, operational expertise and physical location being the big ones.

 

Direct to tape backup is very rare today though, most short term backups are done to disk systems then longer term weekly/monthly copies to tape, Disk to Disk to Tape is what it's know as. You can also make multiple copies of these longer terms backups, one to cloud and one to tape.

 

Home users and business users are so different they should really be covered and considered separately, I mean for home a simple external HDD and a free backup tool is right up there in the best of class methods only really being beaten out by cloud backups. Over engineering solutions is a big problem, but so is under engineering them.

[Bombastinator]

The point about tape made was not that it was useless but that it was old, implying uselessness.  That it is old is true.  That it is no longer useful in many situations is true.  That it is not useful in ALL situations is not true. It’s become a niche thing, and it’s niche is becoming smaller and smaller.

 

Once it was everywhere but that time is gone.
 

[Helly]

Let's analyze what RAID is and what a backup is.

 

RAID = Redundant Array of Independent disks. Nothing in there says backup. Redundancy immediately affects everything connected to it (a file deletion is immediately done on all drives), so undoing this either takes a lot of time or is impossible due to the data being overwritten.

Now as for the definition of RAID being a backup, sure, its a hardware backup. Physical backup of the device storing your data. It prevents immediate data loss due to a drive failure. But is it really? In RAID levels like 1, 5, 6 and 7 (and different level's involving those) the drives are online and spinning and all contain data. They are all needed for the array to be available and normal (not degraded). A physical backup device would be a cold (spare)drive not spinning and not already in the array. So the array itself is NOT a backup and none of the drives in it are either. As all drives in it are used. Hot spares don't count same as cold spares, they are not part of the array and are only pulled in when a drive fails.

 

A backup is simply your file(s) stored on another device that is unaffected by what you do on your local storage. At a later point your data is synced to the backup to apply the changes. It is most likely done with iterations of the file making it possible to roll back any change made to it. Even a deletion. None of this is in any way possible with a RAID array and is all done through software.

 

So in short, RAID is NOT a backup. Unless it is the backup, then its a backup

 

Just thought i'd throw in my 2 cents

[Bombastinator]

8 minutes ago, Helly said:

Let's analyze what RAID is and what a backup is.

 

RAID = Redundant Array of Independent disks. Nothing in there says backup. Redundancy immediately affects everything connected to it (a file deletion is immediately done on all drives), so undoing this either takes a lot of time or is impossible due to the data being overwritten.

Now as for the definition of RAID being a backup, sure, its a hardware backup. Physical backup of the device storing your data. It prevents immediate data loss due to a drive failure. But is it really? In RAID levels like 1, 5, 6 and 7 (and different level's involving those) the drives are online and spinning and all contain data. They are all needed for the array to be available and normal (not degraded). A physical backup device would be a cold (spare)drive not spinning and not already in the array. So the array itself is NOT a backup and none of the drives in it are either. As all drives in it are used. Hot spares don't count same as cold spares, they are not part of the array and are only pulled in when a drive fails.

 

A backup is simply your file(s) stored on another device that is unaffected by what you do on your local storage. At a later point your data is synced to the backup to apply the changes. It is most likely done with iterations of the file making it possible to roll back any change made to it. Even a deletion. None of this is in any way possible with a RAID array and is all done through software.

 

So in short, RAID is NOT a backup. Unless it is the backup, then its a backup

 

Just thought i'd throw in my 2 cents

This is still all about the word “backup” and how it is defined.  I saw the definition “true backup” used by someone earlier.

[Drak3]

7 hours ago, mr moose said:

 

Yes I have, and so have many others:

 

got back to here and start again:

 

 

None of that conflicts with the statement that RAID 1 is a type of backup.

It just explains why RAID 1 isn’t a comprehensive backup.

[WkdPaul]

1 hour ago, Bombastinator said:

This is still all about the word “backup” and how it is defined.  I saw the definition “true backup” used by someone earlier.

I would agree that seems to be the main issue here. Just like people arguing about scientific theories because they apply the general definition of the word "theory" to "scientific theory".

 

Seems to be what's happening here, a redundancy isn't a backup as far as people working in IT are concerned. It's only there to help with downtime, not data recovery.

[Ben17]

On 12/10/2019 at 6:57 PM, dalekphalm said:

By default the circuit is closed. When you hit the switch, you open the circuit, which causes an action.

 

 

It usually open by defualt i think because you need to short the pins with a screwdriver to turn it on

[Mira Yurizaki]

2 hours ago, wkdpaul said:

Seems to be what's happening here, a redundancy isn't a backup as far as people working in IT are concerned. It's only there to help with downtime, not data recovery.

I think it's still a problem that a term is used different contexts, and people going confused cat when people say that X is not Y, even though Y in a different context means X is Y.

[Mira Yurizaki]

Just a random one I like to throw out, which to me at this point is approaching ad-naseum territory. The myth is basically "NVIDIA does not support asynchronous compute"

 

They do, at least since Pascal, because the requirement for asynchronous compute, as defined by AMD themselves is (emphasis added):

Quote

A basic requirement for asynchronous shading is the ability of the GPU to schedule work from multiple queues of different types across the available processing resources.

And then there's this graph showing my GTX 1080 using two different work queues (or engines, as Microsoft calls them)

Spoiler

 

 

[Ben17]

6 minutes ago, Mira Yurizaki said:

Just a random one I like to throw out, which to me at this point is approaching ad-naseum territory. The myth is basically "NVIDIA does not support asynchronous compute"

 

They do, at least since Pascal, because the requirement for asynchronous compute, as defined by AMD themselves is (emphasis added):

And then there's this graph showing my GTX 1080 using two different work queues (or engines, as Microsoft calls them

 

mine has 4 or 5 engines 

[yaya1]

Here's one, hope noone's said it before: Better cooler = lower room temperature

It's obviously not true, because the thermal energy is going to be transferred(with the exception being a cooling system that's not only superior in its cooling capacity but also transfers to somewhere else, kind of like what LTT did some years ago) to the room either way. I know it's pretty obvious, but it's probably just a general misconception, as people think: better cooler = cooler component = less heat = colder room

Just mentioning it because I've seen/heard people say it before...

[Zando_]

3 minutes ago, Jonas_2909 said:

Here's one, hope noone's said it before: Better cooler = lower room temperatur

Oh yeah lol, if anything it makes the room hotter. Due to more efficiently pushing the heat out into the room. Especially noticeable with watercooling loops, they push more heat into the room from the same hardware. Better cooling doesn't magically make your hardware not push as much heat, it just moves it out of the PC faster and more efficiently. 

[Mira Yurizaki]

8 minutes ago, Jonas_2909 said:

Here's one, hope noone's said it before: Better cooler = lower room temperature

I think this is also related to people seeing lower temperatures due to using a better cooler.

 

There's also the idea that say, for example, an AIO is always better than air cooling as far as cooling performance. They go run Cinebench once or maybe prime95 for 5 minutes and say the result to prove their thinking without realizing water has a much higher specific heat than metal.

[Drak3]

48 minutes ago, Mira Yurizaki said:

I think this is also related to people seeing lower temperatures due to using a better cooler.

 

There's also the idea that say, for example, an AIO is always better than air cooling as far as cooling performance. They go run Cinebench once or maybe prime95 for 5 minutes and say the result to prove their thinking without realizing water has a much higher specific heat than metal.

IIRC, JayzTwoCents demonstrated that the whole AiO v high end air cooler debate really hinges on what ambient temperatures are like. AiO tend to outperform high end air coolers in hotter environments, even if the opposite is true in cooler environments.

[leadeater]

8 hours ago, wkdpaul said:

I would agree that seems to be the main issue here. Just like people arguing about scientific theories because they apply the general definition of the word "theory" to "scientific theory".

 

Seems to be what's happening here, a redundancy isn't a backup as far as people working in IT are concerned. It's only there to help with downtime, not data recovery.

Well the thing is backups and the definition of the word in computing is to recover from data loss. RAID is mechanism used to prevent data loss, it cannot be used for recovery when there is data loss. When a RAID 1 array has a disk failure there is no data loss so the mirror disk has not been used to recover from a data loss, it has prevented a loss from happening. The only time there is data loss in relation to disks/hardware in RAID is when the array goes past degraded to failed and when the array is failed you cannot use any part of the array for recovery without sending the hardware array to specialist data recovery services to try and get the data back from your primary and only copy of the data.

 

I don't even see where the confusion comes from in relation to the definition, it's quite clear it's about recovering from data loss or damage so that has to actually happen before you go to data backups, something that does not happen with RAID until it fails or is corrupted.