RE-Evaluating TrueNAS from the Historical Perspective..

[indivision]

Naming things and expiring caches - there is a reason these are frequently named the hardest problems in software development.

As a developer, I would suggest that refers more to naming conventions in the software itself. ie. how to name variables, functions and classes in a way that is helpful.

You create a particular image with some message. And the originator and messenger are the ones whose duty is considering the impact on the target audience and phrasing the message accordingly.

I think this is more of marketing thing than development. But, agreed.

I just don't see where iX has done anything outside of typical use of the terminology.

I'll be giving a talk with much TrueNAS content and endorsement tomorrow and I am still thinking about where to put SCALE in the global context.

If you are interested:

Building infrastructure with Open Source - a TrueNAS and OPNsense walkthrough

t3dd22.typo3.com

Thank you! Looks interesting. I will check it out if I don't end up with a conflict at that time.

 

[NickF]

Hi Again Morgan,

It's the PITA again. :)

I think most of the conflict comes from the expectation that the day 1 use-case was business users with scale-out requirements. That might be your business requirement and our aspiration, but we always understood that production users would be reluctant to leap from Isilon or Nutanix to an unproven SCALE infrastructure. We are planning for a long game, not a short game.

We deliberately chose a strategy to enable home labs and single node systems to be useful to enable a platform of software quality. You can argue with calling it "TrueNAS SCALE" but in the long term, it seems to be a good name and changing names is difficult. Without that early-adopter market, we'd be very slow to deliver enterprise-quality.

The blog for 22.02.2 was actually very clear about the status of TrueNAS SCALE. Some may have missed this, but we try very hard to be clear on where we are at:

I totally understand the strategy, and I'm not arguing that its not a good one. As for being clear as to what IX is communicating the status as being, I whole heartedly disagree and that's the entire point of all of my frustration and the creation of this thread.

I'm not saying that you didn't say what you said or that you were unclear in what you said in those snippets. But let me give you a few more that muddle the waters and make it more confusing.

On the release notes section you outline the "Software Lifecycle". This is a new-ish concept that I am happy to see being used by IX. As a quick side bar, the wording of the "Software Lifecyle", is confusing in an of itself as it uses the TrueNAS CORE nomenclature (RC, RELEASE, U1, U2) and not the TrueNAS Scale nomenclature (22.02, 22.02.1, 22.02.3 etc). But based on that definition, SCALE Angelfish is in U2 (which you yourself also just described it as). Which it defines as "Larger Systems" and "Suitable for higher uptime deployments"

Which is a direct 180 degree difference from what you just quoted from the blog post.

Additionally, lets take a moment to define what TrueNAS SCALE is. I think in order to really get to the heart of this discussion, we need to understand what the actual goal of the project is. We've been talking about this for a while now, and I'm not entirely sure we are on the same page.

In your previous post you stated:

We decided to release this year because it had started to SCALE. At launch in February is did support scale-out in the form of S3 (minio) and Gluster native. We knew that clustered SMB would be available (and wanted) but it needed more testing and polish... we didn't claim clustered SMB at release.

We also identified that there are a large number of users that wanted a single node deployment with both Kubernetes/Docker and a more reliable virtualization environment (KVM). By supporting this single-node model well, we have a much larger user and testing base which delivers quality more quickly. While we all want more automated testing, the reality is that this community gets free software and provides a final and very much appreciated QA role.

Kris Moore stated:

I'd push back against that as well. Angelfish was always intended to be focused on initially porting the "CORE" functionality to its new Linux base, while laying the groundwork for new things like Clustering, Linux Containers, etc. Performance incrementally improved with new updates, and Bluefin is where we start putting SCALE through more of its Enterprise optimization etc. However for most typical community use-cases, you'll find SCALE performance is on par with CORE today. At the higher end, its mixed, better in some cases, worse in others which we are well aware of and are working towards resolving.

As for Clusting specifically, we've even labeled that as early / experimental in the TrueCommand UI, with much more coming as it moves out of "Experimental" phases. But focus has to stay on single node issues first, since that foundation has to be rock solid before you introduce another complex layer like clustering into the mix.

However, if you look at the docs again, you have very clearly stated what TrueNAS Scale is.

TrueNAS SCALE[/B] is the latest member of the TrueNAS family and provides Open Source HyperConverged Infrastructure (HCI) including Linux containers and VMs. TrueNAS SCALE includes the ability to cluster systems and provide scale-out storage with capacities of up to hundreds of Petabytes. Just like TrueNAS CORE, TrueNAS SCALE is designed to be the most secure and efficient solution to managing and sharing data over a network, from smaller home networks “scaled” up to massive business environments.

SCALE is an acronym that represents the core features of the software: [LIST] [*][B]S[/B]caled-Out ZFS [*][B]C[/B]onverged [*][B]A[/B]ctive-Active [*][B]L[/B]inux Containers [*][B]E[/B]asy to Manage [/LIST]

Let's take a look at what we have in Angelfish and compare to what the definition of your product is.

• • Open Source - YES
    • Defined by you as: Free to download and use, TrueNAS SCALE welcomes developers and testers to contribute to its Open Source development model.
  • Scale Out - Barely a yes, you can do this, there aren't really any use cases that can leverage this feature yet, and Clustered SMB was only released 2 days again and breaks normal SMB shares.
    • Defined by you as: OpenZFS and Gluster combine to enable scale-out ZFS capabilities with excellent stability and very efficient compression and snapshots.
  • Hyperconverged- Not by your own definition!
    • Deploy a single hyperconverged node in a home/office or a cluster with hundreds of compute and storage nodes in a datacenter.
  • Virtualization - YES
    • With support for KVM VMs, Kubernetes, and Docker containers, it’s easy to add applications to suit your every need.

The docs go further:
Unlike other HCI platforms, a user can get started with TrueNAS SCALE on a single node and incrementally scale up and scale out to over 100 storage nodes with many additional compute-only nodes. TrueNAS SCALE is true Disaggregated HCI, meaning storage and compute can be scaled independently. Each node can support Virtual Machines (with the KVM hypervisor) as well as Docker containers by using native Kubernetes.

The definition of SCALE, the end goal of what SCALE will be is clearly defined. But Angelfish U2 != SCALE. It's a baseline. It's a subset of features. Its pretty good, despite some REALLY stupid bugs like VMs not being able to talk to SMB shares on the host they live in. But my gripe isn't that the current state of SCALE is bad, but it's marketing is confusing. It's goals are not met, but we are not in a beta.

The closest analogy to another piece of software I can think of is Star Citizen. That game has been in development for eons, but it's making money hand over fist. It's in a perpetual state of alpha, it has HUGE lofty goals that aren't met. But that's okay, because the content that's there is good enough for now and they aren't kidding themselves by saying the game is "released".

EDIT: Its sort of like you guys are speaking out of both sides of your mouth. On one hand, you clearly define what SCALE as a product represents. On the other, you "release" a product, and define it as product in U2 which is "Suitable for higher uptime deployments", but that "released" product does not live up to the very definitions you gave to your own product. Angelfish is a milestone, not a finished product. SCALE, as an idea, as a whole, is still under development. It really should be labeled as a beta or a technical preview.

 

[danb35]

I recall the main reason given by iX was that there was too much of a gulf between the two code-bases

I'm sorry, but your recollection is incorrect. See https://www.truenas.com/community/threads/important-announcement-regarding-freenas-corral.53502/:

we’ve also seen nearly half of the initial users revert back to FreeNAS 9.10. User feedback about this drop-off has been clear: challenges upgrading from 9.10, general instability, lack of feature parity with 9.10 (Jails, iSCSI, etc), and some users experiencing lower performance than expected given the increased demands FreeNAS Corral has on system hardware resources.
...
In doing so, we discovered some holes in the architecture which make enterprise-quality file access using 9pfs impossible without a lot more effort and soak time, prompting us to to also re-think how to more safely enable this capability.
...
we have decided that the amount of work still required to bring FreeNAS Corral (as currently architected) up to an acceptable standard for quality, reliability, and data integrity will take an unreasonable amount of time.

IOW, what they foolishly decided to call Corral was a raging dumpster fire, beyond economical repair.

Why bring it up? Because some of us have long memories. Because iX has screwed the pooch before. Because they promised us they'd do better. And though they haven't quite sunk to that level since, there have been some pretty major blunders. And because, as I observed up-thread, SCALE really does seem to be, in its way, another attempt at what FN10 was trying to do.

 

[NickF]

I'm sorry, but your recollection is incorrect. See https://www.truenas.com/community/threads/important-announcement-regarding-freenas-corral.53502/:

IOW, what they foolishly decided to call Corral was a raging dumpster fire, beyond economical repair.

Why bring it up? Because some of us have long memories. Because iX has screwed the pooch before. Because they promised us they'd do better. And though they haven't quite sunk to that level since, there have been some pretty major blunders. And because, as I observed up-thread, SCALE really does seem to be, in its way, another attempt at what FN10 was trying to do.

I agree, and that is VERY MUCH why I drew the initial correlation between the two. It's difficult to harvest the pertitent details (beyond my own recollection and some posts and news articles) of exactly what Corral aimed to be because IX has buried a lot of that information. But terms like hyperconverged, Docker and virtual machines were all part of the Corral goal. Arguably, I think I liked the interface for the docker implementation a little bit more than what we have now with TrueCharts. It felt like what the industry was already doing with Portainer, rather than recreating the wheel. Alot of that was eventually ported into the FreeNAS 11 codebase, and SCALE is certainly the next logical step in that evolution. Obviously technology has progressed a bit since then, and IX alot of lost engineering hours that went into a codebase that was Corral and re-doing alot of that work in 9.10/11.

Interestingly, Jordan Hubbard himself replied to the reddit cross-post:

Reddit - Dive into anything

www.reddit.com

 

[jgreco]

Yeesh.

 

[Samuel Tai]

As the participants are just repeating prior arguments, let's just agree to disagree. I'm closing this thread for further comments.

 

[Ericloewe]

Just a note: If anyone has something new to add, just report this and ask for assistance.