RAIDZ expansion, it's happening ... someday!

Constantin

Vampire Pig
Joined
May 19, 2017
Messages
1,828
Everything has a place. Some of my friends may be extremely accomplished professionals but they have essentially zero interest in maintaining an IT infrastructure. Thus, there always will be a market for very simple, adequately-performing storage - either on-site or in the cloud. That’s Synologys target and the hardware works well enough for that purpose.

Where people can get into trouble is the unicorn solutions that the likes of QNAP, ReadyNAS, and Synology occasionally come up with such as the “DAS/NAS” that QNAP once marketed. It’s not a DAS using Thunderbolt, it’s Ethernet over Thunderbolt and any device on the Thunderbolt bus is taken over by the QNAP NAS by default. That’s a problem if you have a computer with just one Thunderbolt port and would like to use a monitor also (Hello, MacBook Air).

so the cloud has its place and its yet another tool to use and better than nothing when it comes to off-site backups.
 

HarambeLives

Contributor
Joined
Jul 19, 2021
Messages
153
better than nothing when it comes to off-site backups.

You say that as if "the cloud" is one thing, and is not really a great option for off-site backups

How many people are running geo-redundant, highly available arrays for off-site backups? Probably only the people using the cloud
 

ChrisRJ

Wizard
Joined
Oct 23, 2020
Messages
1,909
In my case "the cloud" comes with the challenge of bandwidth. Living in a somewhat rural area in Germany means to only have an upstream capacity of around 20 Mbps for me. So just uploading the most critical data took me around 6 months.
 

Constantin

Vampire Pig
Joined
May 19, 2017
Messages
1,828
For some, having an automatic iCloud backup as part of their MacOS or iOS experience is perfect - its not cheap, it’s not fast, but it works for them well enough.

offsite storage that is fast, redundant, and reliable tends to be a business product and priced accordingly. Some companies even allow you to send an initial backup via disk, and they can send your backups on disks back if you have a catastrophic event.

all comes down to what you want to pay and how much time you want to put in. And if you have friends or family that are like-minded and that you trust, it’s even possible to do offsite backups to each other’s NAS’ of critical content.
 

Constantin

Vampire Pig
Joined
May 19, 2017
Messages
1,828
The big delta re: @jgreco’s data vs. that stored on cloud providers hardware is that he knows where it is, what health it’s data storage system is enjoying, and so on.

Most users of backblaze, iCloud, AWS, or Google users have no idea what it takes to keep data safe, they just want a reliable, inexpensive service to do it for them. Backblaze may publish a number with a lot of nines but none of that is relevant when among the many petabytes under management, it’s your stuff that just went into the bitbucket.

bottom line, it comes down to preferences and how important it is to you to have access to the data quickly. To me, the cloud is a useful tool but it shouldn’t be relied on exclusively. A backup solution with single point of failure is never a good idea unless you don’t care about the data - but if that’s the case, why pay $ to back up?

Fast local backups to DAS / NAS combined with the cloud is likely a sweet spot for many SOHO users. It allows quick recovery from “normal” meltdowns while even slow offsite backups allow eventual data recovery when truly unhappy stuff occurs locally.
 

jgreco

Resident Grinch
Joined
May 29, 2011
Messages
18,681
One of the more interesting bits of "cloud" is that Apple iCloud is back-ended by AWS/Google/Azure/etc. This might actually be great if your data was being securely stored on multiple clouds in an encrypted form. Quite frankly, that'd probably be more reliable than anything you or I could manage individually.

The problem is that these are all for-profit companies, and the cost in the long run is a bit pricey. I just got a quote from a cloud vendor a month ago asking for $1758/month for 14TB of VM storage space backed by HDD. AWS wants $2/TB (or $28/14TB/mo) for general storage, and that actually seems quite reasonable to me, except that you get terrible transit bandwidth fees...

By way of comparison, I can get a really decent used rackmount server for $1K, four 14TB HDD's for $400/ea, and run FreeNAS for $2600, over something like a five to ten year lifespan. Hell, let's be really aggressive, a 25% occupancy rate and a high end $2K FreeNAS server, and twelve 14TB drives to have four spares available, that's a $6800 server, and let's say it only lasts three years, that's $189/month amortized over 3 years.

So here's the funny bit. You can get a full rack and 50Mbps of bandwidth at ServerCentral, now Deft, for $1500/month. (Please note that I am absolutely connected to these people, but my estimation that they are high quality is independent of that fact). So you can get a full rack and reasonable bandwidth and a whole server for $1689/month, plus this gives you a place to put all your other servers too.

You can find much cheaper options than all of this, by the way.
 

Ithaca

Cadet
Joined
Aug 18, 2014
Messages
3

The raidz expansion is getting ready for use in FreeBSD. Any news on the progress or implementation on truenas?
 

jgreco

Resident Grinch
Joined
May 29, 2011
Messages
18,681
Yeah. First reported in 2017, see first post in this thread, I expect it'll happen sometime in the next five years or maybe a year or two after that. :smile: Kidding. (Am I?)

You probably want to read a little closer what the article actually says, which includes the words "not yet integrated". We've heard lots of "later this year" promises for ZFS features in the past, which ended up being pushed off for a long time, so don't expect that this is going to hit TrueNAS this year even if it somehow manages to squeak into FreeBSD by the end of the year. It is a big, somewhat dangerous feature and there is likely to be some caution and paranoia before integrating it into TrueNAS or making it available to users.
 

Yorick

Wizard
Joined
Nov 4, 2018
Messages
1,912
“We aim for this to be integrated by Q3.” - as the eternal optimist I’ll say it’s in FreeBSD by 2023 and in TrueNAS by 2024. I know, I know, that’s practically tomorrow!
 

Yorick

Wizard
Joined
Nov 4, 2018
Messages
1,912
"The Moore" just dropped this comment in the PR: "FWIW, we on the iX side are taking a serious look at this right now to determine when to incorporate it into TrueNAS SCALE. Once we have something ready for testing we'll be sure to ping back in case folks want to help us beat it up and get it into shape for inclusion into upstream."
 

Etorix

Wizard
Joined
Dec 30, 2020
Messages
2,115
It will be integrated by Q3 then. Just do NOT specify "Q3" of which year…
For what it's worth, I note that "The Merge" (Ethereum) occurred first.
 

Ericloewe

Server Wrangler
Moderator
Joined
Feb 15, 2014
Messages
20,176
I half-expect someone to figure out a reasonable BPR before this shows up in production.
 

Davvo

MVP
Joined
Jul 12, 2022
Messages
3,151

Ithaca

Cadet
Joined
Aug 18, 2014
Messages
3
i had a little hope for the openzfs summit of 2022 to hear more about the status of the raidz expansion. As far as I read about it, it wasn’t even mentioned.
 

garm

Wizard
Joined
Aug 19, 2017
Messages
1,555
I am terrified about the post expansion era… the flood of lost pools that will kill off consumer confidence in ZFS
 

Whattteva

Wizard
Joined
Mar 5, 2013
Messages
1,824
I am terrified about the post expansion era… the flood of lost pools that will kill off consumer confidence in ZFS
Yeah, this!
Oh well, I don't really care cause I run striped mirrors.
 

sretalla

Powered by Neutrality
Moderator
Joined
Jan 1, 2016
Messages
9,702
I am terrified about the post expansion era… the flood of lost pools that will kill off consumer confidence in ZFS
I wouldn't be too worried about that. The expansion process isn't actually that complicated or bad (doesn't even move much of the existing data anywhere... it just remains striped as before with the old parity).

As long as there aren't any major bugs in something like scrub code, I don't see why anyone would lose a pool in doing it.

I guess something like the rebalance script is something that may become popular after running the expansions... https://github.com/markusressel/zfs-inplace-rebalancing

(personal note: I don't like the idea of having a pool that was expanded that way myself, but there's actually nothing bad about it per se).
 
Top