Goodbye FreeNAS

[danb35]

That's also true. "Bullet proof", whether literally or figuratively, is always relative--there's always a bigger/faster bullet.

 

[Stux]

What you need is data integrity control/detection.

It's no good having a backup if you don't know when to restore from it, or you replace the backup with corrupt data.

 

[garym]

While I agree with your sentiment, there is no such thing as a "bullet proof NAS". People should be following the 3-2-1 rule for truly irreplaceable data. Raid is not a backup and disasters happen.

you are right, but if one dosen't make an effort to reach it , bullet proof, then there is no way one will get close. Raid is a method not an object, A NAS device is an object of storage and can very well be a backup device. A backup that is supported by the best file system available to us today, ZFS. I use just such a system for backups and when not in direct use is offline. ZFS gives me clean data and when offline, the best security.
The other part of bullet proof is the ability to replace a disk without causing more damage. A good NAS will show you which disk to change, not leaving you to guess or having to label each disk as you add them to your NAS. Alternatively you could have a menu selection to ID any drive, or be able to use the dd command. directly.

 

[danb35]

A good NAS will show you which disk to change,

...which is a great idea, and trivially easy if the software vendor also controls the hardware (see, e.g., TrueNAS). But if the software runs on arbitrary hardware, much more difficult.

 

[danb35]

FreeNAS 11 has had a much more rational development cycle.

In light of iX' bizarre decision to put 11-RC into the newly-created 11-STABLE train, do you still hold this position?

 

[danb35]

Raid is a method not an object, A NAS device is an object of storage and can very well be a backup device.

Well, yes. When you analyze the phrase "RAID is not a backup" closely, like many clever phrases, it doesn't hold together too well. But the meaning is, I think, fairly clear--storing your data on a RAID does not replace backing it up. RAID will protect against disk failure consistent with its redundancy, but it won't protect against (for example) catastrophic hardware failure, natural disaster (which is why I made a trip to Atlanta with a laundry basket full of hard drives about six months ago), malware, or user error. The latter two risks can be mitigated by a good snapshot strategy, but that's outside the scope of RAID as such.

That's the point of that saying: If your data's important, it should be backed up on another machine/device, and ideally offsite (which is, of course, much easier than in the days before the Internet). Putting it on a RAID array does not satisfy the requirement to back it up. If your only copy of your important data is on your FreeNAS box, you're in a vulnerable situation, no matter how good your hardware and how redundant your RAID configuration.

 

[fracai]

In light of iX' bizarre decision to put 11-RC into the newly-created 11-STABLE train, do you still hold this position?

From my observation it has indeed been a much saner development process. For one, it's not throwing away jail compatibility, which means there will actually be a transition path that doesn't break things. Also, while there is a new GUI available, the old one is still there as well. In generall 11 is making smaller incremental changes towards the same goals. That does seem much saner than what Corral brought.

And finally, let's not forget that one of the two hard problems in Computer Science is naming things. I'm sure the train that would lead to a stable 11 release seemed sensible to be named as such, ignoring the implication that it had already arrived there.

 

[Ericloewe]

In light of iX' bizarre decision to put 11-RC into the newly-created 11-STABLE train, do you still hold this position?

Yes, it's an unfortunate name choice, but the development itself is sane.

 

[devster]

I'm probably going a little opposite-over-adjacent, but I'm also baffled that Rockstor is using btrfs. The biggest advantage of btrfs, IIRC, is that you can add drives "safely"--that is, you can turn a three-disk RAID5 into a four-disk RAID5. That is pretty big, especially for the home market, and it's something that ZFS will probably never see. But other than that, I can't imagine why anyone would prefer btrfs over ZFS.

If I understand correctly you can generally rebalance the array online, for example switching from raid 10 (supposedly stable), to raid 6 once they fix it. If it manages to deliver on reliability I think this would be a killer feature for home use.

Sent from my ONEPLUS A3003 using Tapatalk

 

[danb35]

If it manages to deliver on reliability I think this would be a killer feature for home use.

Indeed, but "if it manages to deliver on reliability" is critical--and the filesystem is ten years old, and they don't sound like they're anywhere close with the RAID5/6 code yet, and they also have made some really stupid design decisions along the way (like no checksums for parity data, and writing in a write hole). But yes, the ability to shift RAID levels, add and remove disks at will, would be huge for the smaller user.

 

[CompuGlobalHyperMegaNet]

RAID will protect against disk failure consistent with its redundancy, but it won't protect against (for example) catastrophic hardware failure, natural disaster (which is why I made a trip to Atlanta with a laundry basket full of hard drives about six months ago)

Even though technology has progressed so far in the last few decades, there's still no beating the bandwidth of a station wagon (or laundry basket), fully loaded with drives, hurtling down the freeway.

---------------------------------------

I'd just like to put in my two cents on the whole ZFS vs. BTRFS and FreeNAS vs. WhateverOS thing. For me, it comes down to three simple things.

1. The question of whether data integrity is the number one deciding factor in choosing an OS.
2. FreeNAS uses ZFS because of it's (pretty much) unparalleled data integrity features.
3. BTFRS is simply not stable or production ready.

Surely that can only lead to one conclusion... if data integrity is of the utmost concern, then ZFS is the clear choice (either on FreeNAS, *BSD or Linux). Where as if data integrity isn't a top priority, then (as much as I love it) FreeNAS probably isn't the best choice and BTRFS is still a bad choice considering that there are other, much more mature filesystems out there.

 

[danb35]

no beating the bandwidth of a station wagon (or laundry basket), fully loaded with drives, hurtling down the freeway.

The laundry basket was in a station wagon, at least until the station wagon was wrecked. Then it was in a mini-van. But, even though the bandwidth is great, the latency sucks. It beats IP over Avian Carrier, however.

 

[Ericloewe]

It beats IP over Avian Carrier, however.

That's because you haven't tried African swallows yet. I hear they're rather fast when not loaded with a coconut.

 

[danb35]

That's because you haven't tried African swallows yet.

They're kind of hard to source around here. I'd expect it was easier in England around 932 AD.

 

[Jailer]

That's because you haven't tried African swallows yet. I hear they're rather fast when not loaded with a coconut.

African or European?

 

[Ericloewe]

They're kind of hard to source around here. I'd expect it was easier in England around 932 AD.

Doubt it, they're non-migratory.

 

[NZ_JJ]

Doubt it, they're non-migratory.

I'd just load up a cow on a catapult


Sent from my iPhone using Tapatalk

 

[Stux]

I'd just load up a cow on a catapult


Sent from my iPhone using Tapatalk

Is that a Corral joke?

 

[Jailer]

 

[Arwen]

The laundry basket was in a station wagon, at least until the station wagon was wrecked. Then it was in a mini-van. But, even though the bandwidth is great, the latency sucks. It beats IP over Avian Carrier, however.

I heard some company was actually using IP over Avian Carrier. Basically they took digitial pictures of campers / rafters. But, they wanted the pictures printed out and ready for the customers when they checked out. So they sent a flash drive with photographs on a homing pidgeon. If I recall correctly, this was north of Boulder, Colorado, a few years ago.

Edit: Oops, it was Other avian data transfer methods;

https://en.wikipedia.org/wiki/IP_over_Avian_Carriers#Other_avian_data_transfer_methods
http://origin.denverpost.com/opinion/ci_6209735

And more than 10 years ago.