MKBHD builds a 140TB unRaid Server


Kewjoe

Recommended Posts

Ouch.. 1 parity for 14 drives all bought at the same time? Double ouch! And they called it backed up? Any raid is never a backup, additional redundancy yes but never a backup.

I have seen some other stuff various boards posted about the Linus guy that will make you shudder dealing with hard drive arrays.

Link to comment

Yes I was confused when we said that was safe having all those drives with only 1 parity drive! Maybe they thought mentioning 2 parity drives would confused people? Anyway, the ad, I mean video was clearly paid for by Seagate and the company that made the chassis so I'm sure they don't really care.

Link to comment

Linus is famous for knowing very little making stupid decisions (like running his main server in a crazy undafe RAID config with no backups which he then lost), he is entertaining which is why he's popular. These guys get all kinds of hardware for free so companies can get some free advertising, meanwhile a lot of people with real content and knowledge are unknown. Its the culture we live in.

Link to comment
8 hours ago, digiblur said:

Ouch.. 1 parity for 14 drives

They did explain that due to how unRaid works without striping that they can always take advantage of Seagates included data recovery service at no charge during the warranty period.

 

That, and unlike a traditional RAID system, a URE does not constitute a failure of a drive within unRaid.  Pure RAID systems value data integrity above all else.  Its better to lose ALL of the data than have any part of it compromised.  This is what you want you bank to run

 

unRaid values resiliency.  The odds of having any data compromsed is small but does exist, and in the event of a situation where data may get compromised, then you're better off having most of it (maybe only a single file got trashed) than none of it.  This is what I want to run.

Edited by Squid
Link to comment
5 hours ago, digiblur said:

Fully aware of how unRaid works but no way would you catch me running 14:1 on any type of array with the same drives bought at the same time.

 

These are probably the most reliable drives you can buy, and they have data recovery service on these.

And of course they didn't pay for any of this. I wish they'd covered more details.

 

Link to comment
 
These are probably the most reliable drives you can buy, and they have data recovery service on these.
And of course they didn't pay for any of this. I wish they'd covered more details.
 


Dumb decision no matter what drives were used. Linus is known to do some dumb things with arrays so this is the norm for him.
Link to comment
Fully aware of how unRaid works but no way would you catch me running 14:1 on any type of array with the same drives bought at the same time.

I, and I assume many others, have never seen any evidence that drives from the same batch are more likely to die together than drives from separate batches.

As to your example of nobody would every catch you running 14:1, I ran (4) chassis all running 28:0, no parity drives at all. Parity isn't a backup method, never has been, and never will be. My chassis ran along just fine for years. Two were main servers and two were backup servers containing an identical set of data disconnected from power and network.

Since unRAID doesn't stripe data, 14:0, 14:1, or 14:2 is honestly just personal preference. Unless you have at least one full backup of your data, and preferable a third set off site, your playing with fire.


Sent from my iPhone using Tapatalk
Link to comment
4 minutes ago, CyberSkulls said:

drives from the same batch are more likely to die together than drives from separate batches

Handling issues. If that particular box of drives got dropped somewhere in transit, chances are good they all suffered the same trauma. I see many more failures from careless handling than from manufacturing defects.

  • Upvote 1
Link to comment
Handling issues. If that particular box of drives got dropped somewhere in transit, chances are good they all suffered the same trauma. I see many more failures from careless handling than from manufacturing defects.

Now that I would defiantly agree with. Amazon/Newegg craptastic packaging FTW!!


Sent from my iPhone using Tapatalk
Link to comment

With that much data does it really matter if its 1 or 2 parity drives? You need to have backups in any case. There are plenty of users with >15 drives an donly 1 parity drive, 2 is a very new feature. And the drives were not the same batch, they got it in 2 shipments. Even if they were I very much doubt it makes any difference. In fact if all of them got damaged thats a good thing as you'd be able to replace them immediately.

Link to comment
18 minutes ago, MrCrispy said:

In fact if all of them got damaged thats a good thing as you'd be able to replace them immediately.

Damage doesn't always show up until later, and doesn't always effect all drives in a shipment.  That's one of the biggest reasons to run several preclear cycles, to increase confidence.

Link to comment

I, and I assume many others, have never seen any evidence that drives from the same batch are more likely to die together than drives from separate batches.

As to your example of nobody would every catch you running 14:1, I ran (4) chassis all running 28:0, no parity drives at all. Parity isn't a backup method, never has been, and never will be. My chassis ran along just fine for years. Two were main servers and two were backup servers containing an identical set of data disconnected from power and network.

Since unRAID doesn't stripe data, 14:0, 14:1, or 14:2 is honestly just personal preference. Unless you have at least one full backup of your data, and preferable a third set off site, your playing with fire.


Sent from my iPhone using Tapatalk


Why are we all here?
Link to comment
57 minutes ago, digiblur said:

 


Why would I post a question to myself on a forum to others... That would be weird.

 

 

Why are you being so aggressive? You use unRaid how you like, let others use it how they like? What's so difficult about that concept?

Link to comment
11 hours ago, CyberSkulls said:


I, and I assume many others, have never seen any evidence that drives from the same batch are more likely to die together than drives from separate batches.

As to your example of nobody would every catch you running 14:1, I ran (4) chassis all running 28:0, no parity drives at all. Parity isn't a backup method, never has been, and never will be. My chassis ran along just fine for years. Two were main servers and two were backup servers containing an identical set of data disconnected from power and network.

Since unRAID doesn't stripe data, 14:0, 14:1, or 14:2 is honestly just personal preference. Unless you have at least one full backup of your data, and preferable a third set off site, your playing with fire.


Sent from my iPhone using Tapatalk

 

Agree with everything. If I have a backup server with frequent rsyncs, parity is not that essential and may in fact be not preferred. Its the same tradeoff you make when choosing to have a non-mirrored cache, which many people also have.

 

I know everyone says you need 3 backups with 1 offsite, but in practice very few can afford that with large datasets.

Link to comment

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.

Guest
Reply to this topic...

×   Pasted as rich text.   Restore formatting

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.