TrueNAS might not be for you, if you are home user.

[morganL]

This is a moot point over a trusted local network. NFS is "insecure" for the same reasons. Yet using rsyncd and NFS over a local network is not considered dangerous, and is fully supported.

Rather than remove Rsyncd + Modules, why not leave it in, but warn the user with a disclaimer that it uses a non-encrypted stream and has no robust user authentication? (The same warning message can be applied to NFS as well.)

Without specifying IP/host whitelists, any client can connect to an NFS share or Rsync Module. This has always been true for a local network; but no one's saying to remove NFS. So why is Rsyncd being singled out?

About 20% of users use NFS. It is primarily used on LANs.

A very small % use rsyncd .. and it can be moved to an App. There is no restriction on the rsyncd apps.

We did leave rsync on ssh in the system.

 

[winnielinnie]

and it can be moved to an App.

What about its fate for Core?

 

[Whattteva]

Last few years, and especially with transition to TrueNAS Scale, I couldn't notice that I am not the target customer of the developers of this system. Which makes sense, I don't pay anything. iXsystems is making money from enterprise selling them their enterprise solutions based on TrueNAS. But this created misalignment in incentives between us - home users and the developers of this system.

That's interesting. I happen to think the other way. SCALE, to me, is more geared towards home users, especially with the whole Apps section. I'd imagine Enterprise users care more about stability over flashy new features (at least with all the companies I've worked for so far). We were running Windows XP for like at least 2-3 years after Windows 7 was released cause IT department had to test Windows 7 thoroughly.

I personally couldn't care less for the Apps thing and would rather have the rock solid stability and better memory management of FreeBSD for ARC, so I have little use for SCALE and only run CORE.

I really believe TrueNAS Core (and possibly SCALE) would be better served with a clear distinct "Community" edition.

CORE and SCALE are kind of the "community" edition. The enterprise version is obviously TrueNAS Enterprise.

ALSO, take a page outta the documentation from the folks over at netgate.. absolutely close to, if not, THE BEST documentation i've ever come across... it is so good i've been able to use their documentation on configuring other devices.. just my .02

Netgate is even worse IMO. I switched to OPNsense due to this bug which people has suffered since 2017!!! with no end in sight. They put the blame on Unbound, but OPNsense (uses the same Unbound DNS service) exhibits no such issue with the same settings (ad-blocker and DHCP DNS registration).

I don't think RedHat is a positive example of anything right now.

I'd argue even long in the past. The systemd debacle, in particular, pissed Linus Torvalds so much that he banned all code from one of the main developers (Kay Sievers) who also happens to work for RedHat.

About 20% of users use NFS. It is primarily used on LANs.
A very small % use rsyncd .. and it can be moved to an App. There is no restriction on the rsyncd apps.

Sounds about right. I myself only use NFS and SMB. I really only use rsync for ad-hoc CLI copying that's more robust than the shell's built-in cp. In my opinion, rsync is kinda' obsolete compared to zfs send/recv for regularly scheduled backups. Of course, I'm probably a minority since I have 2 nodes that both run ZFS, which enables me to use zfs send/recv.

 

[winnielinnie]

CORE and SCALE are kind of the "community" edition. The enterprise version is obviously TrueNAS Enterprise.

Core is the codebase for Enterprise. As far as I know, this started after "FreeNAS" became "TrueNAS Core". Other than branding, TrueNAS Core is essentially the same software as Enterprise.


When I say "Community" edition, I mean in the sense of rapid development and new features. A "playground" of sorts, in which the enterprise customers will be unaffected. Eventually, if so desired, mature and stable features can be included in the enterprise edition.

Just using TrueNAS Core, I have a long list of bugs and features that would greatly enhance the usability and experience. It's a "bam, bam, bam, bam" type of mental process. It would be an uphill battle to make a bug report and feature request for each and every one of them, because there's always the "How does this benefit our enterprise customers?" angle, which meets many requests with friction and resistance. (Not to mention the fact that SCALE benefits from Core feature requests, since Core is being treated with a "feature freeze" approach, akin to Ubuntu LTS.)

A distinctly separate community codebase wouldn't be so resistant to quick enhancements and new features.

 

[morganL]

What about its fate for Core?

No current plans for CORE to change in this regard.

 

[morganL]

Core is the codebase for Enterprise. As far as I know, this started after "FreeNAS" became "TrueNAS Core". Other than branding, TrueNAS Core is essentially the same software as Enterprise.


When I say "Community" edition, I mean in the sense of rapid development and new features. A "playground" of sorts, in which the enterprise customers will be unaffected. Eventually, if so desired, mature and stable features can be included in the enterprise edition.

Just using TrueNAS Core, I have a long list of bugs and features that would greatly enhance the usability and experience. It's a "bam, bam, bam, bam" type of mental process. It would be an uphill battle to make a bug report and feature request for each and every one of them, because there's always the "How does this benefit our enterprise customers?" angle, which meets many requests with friction and resistance. (Not to mention the fact that SCALE benefits from Core feature requests, since Core is being treated with a "feature freeze" approach, akin to Ubuntu LTS.)

A distinctly separate community codebase wouldn't be so resistant to quick enhancements and new features.

By all means, start a thread with all of the bugs and features you'd like done.

The test is not "how does it benefit Enterprise customers.?" but "Does it negatively impact Enterprise customers?".

Then the other question is who is doing the development and how is that being funded? We don't charge for consumer users... so we are looking for volunteers or a funding source.

 

[Hello_World]

By all means, start a thread with all of the bugs and features you'd like done.

The test is not "how does it benefit Enterprise customers.?" but "Does it negatively impact Enterprise customers?".

Then the other question is who is doing the development and how is that being funded? We don't charge for consumer users... so we are looking for volunteers or a funding source.

Why not add a TrueNAS SCALE Pro version for geek users, provide enhanced functions closer to home users, and charge a certain fee with reference to Unraid.

Synology bundles DSM's system and hardware to provide high-quality and out-of-the-box NAS services, and thus has a high market share in the NAS system market.

Each user actually paid over $100 for the DSM system. Unraid offers licenses for $59/89/129 with the ability to get bulk user purchases. Clearly, Synology and Unraid gained high visibility through the consumer market rather than the enterprise market. For the most part, these companies are profitable even without considering enterprise customers.

TrueNAS is entirely aimed at the enterprise market. The addition of most new functions only considers the needs of enterprises, which greatly increases the cost of general users, but there is no way to obtain sales revenue from general users.

In the community, we can see such a scenario: many users' needs cannot be officially responded to; official personnel need to spend a lot of effort to reply and investigate problems that would not arise in an enterprise scenario.

TrueNAS is the only NAS system I've come across that requires a lot of time to follow the community and news. For many users, this is hopeless.

iX needs to adjust its product line and redefine its relationship with general users.

Perhaps we can conduct a large-scale survey and user interviews to assess the possibility of further improvement. For users, paying an appropriate amount for an excellent product is acceptable in most cases.

 

[ChrisRJ]

Upfront disclaimer: I obviously cannot speak for iXsystems, since I am not an employee or associated in any other form.

If the idea is that paying around 100-200 USD/EUR for TrueNAS Core as a hobbyist user per year, this is not necessarily something that is attractive for a company in the enterprise business. And that is under the assumption that no support comes with that payment. It is certainly not attractive if support comes into play.

 

[Johnny Fartpants]

Personally I don't put TrueNAS in the same box as Synology. In the home-lab world then sure but in a large enterprise environment I don't see them operating but I do see TrueNAS. As a result Synology seem to target their development towards ease of use and home-lab / small business style deployments. No doubt iX would like a bigger piece of that pie and the mini series coupled with SCALE I suppose is trying to make that more compelling but I think their target audience is very different once you get passed the fact they are all NAS. I wouldn't deploy Synology at scale in enterprise but equally I wouldn't recommend my non-technical friend buy a TrueNAS Mini but I may recommend they get a Synology.

Companies that have stood the test of time have often had one thing in common and that is they have stayed focused at what they are good at and tried not to be all things to all people. Don't get me wrong you can't stand still either and rest on your laurels but staying focused on what you do good is key.

Steve Jobs once said that "Focusing is about saying no".

 

[winnielinnie]

By all means, start a thread with all of the bugs and features you'd like done.

I've thought about doing that, and I still may do so. But you can understand my tentativeness? From my experience on the bug tracker with the resistance and friction, it saps my motivation.

As an example: the Shell feature is broken for Core (an "enterprise-quality" product) is going to be fixed only for SCALE... from a Core bug report. That hurts.

 

[danb35]

the Shell feature is broken for Core (an "enterprise-quality" product) is going to be fixed only for SCALE... from a Core bug report.

...and this is something that's been broken for many (over ten) years, with many complaints.

 

[Johnny Fartpants]

As an example: the Shell feature is broken for Core (an "enterprise-quality" product) is going to be fixed only for SCALE... from a Core bug report. That hurts.

This is a VERY good point. I too raised a bug with Core and it only got fixed in SCALE.

 

[morganL]

I've thought about doing that, and I still may do so. But you can understand my tentativeness? From my experience on the bug tracker with the resistance and friction, it saps my motivation.

As an example: the Shell feature is broken for Core (an "enterprise-quality" product) is going to be fixed only for SCALE... from a Core bug report. That hurts.

What was the NAS-Ticket?

When a module needs a major rewrite or new UI, we have been favoring TrueNAS SCALE.
If it's bug fix... we should do it in both if applicable.

There will be a TrueNAS CORE 13.1 in 2024 for major ZFS and SMB updates.

 

[winnielinnie]

What was the NAS-Ticket?

[NAS-104469] - iXsystems TrueNAS Jira

ixsystems.atlassian.net

Re-attempted to bring attention to Core, and immediately shut down for the same reasons given in the above ticket:

[NAS-120262] - iXsystems TrueNAS Jira

ixsystems.atlassian.net

 

[morganL]

[NAS-104469] - iXsystems TrueNAS Jira

ixsystems.atlassian.net

Re-attempted to bring attention to Core, and immediately shut down for the same reasons given in the above ticket:

[NAS-120262] - iXsystems TrueNAS Jira

ixsystems.atlassian.net

The developers found it required a major change and there was risk involved.
Is there no workaround that is acceptable?

If this is a high priority for many users, I can ask for another review.

 

[winnielinnie]

If this is a high priority for many users, I can ask for another review.

How isn't it? The "Shell" is nearly useless as a feature in Core, and you end up using SSH in a terminal anyways.

The way it's presented: It's a quick and convenient means to use the command-line while you're logged into the NAS appliance's web GUI.

Being unable to properly copy text is already bad enough. (It copies text from an area away from your highlighted selection.) But on top of that, the text wraps around way too early. It cuts and wraps only a third of the viewable screen space. So the "Shell" feature simply isn't worth it. It's a broken feature that's remained broken, and you begin to wonder why it wasn't outright removed with a message "Setup SSH, and use your client's terminal to use the command-line."

While you're in the web GUI, being able to quickly use a fully-fledged command-line shell keeps your activity contained within the appliance's GUI. Quickly checks logs, ARC info, ZFS commands, folder lists, etc, etc, etc, and being able to copy + paste the text that you select.

 

[Patrick_3000]

I'm a home user, and I used Freenas for many years, then switched to Truenas Core, then last year switched to Truenas Scale.

I can say categorically that I made the right decision to switch to Scale. Core is fine, but my guess is that it won't be maintained over the long run because it's not useful to the enterprise and to iXsystems' business model.

Also, Linux Debian, the base of Scale, is compatible with a far greater selection of hardware than FreeBSD, the base of Core. This is not just theoretical. At one point, I was using a 2.5GBPS network card that was recognized by Scale but not Core.

Moreover, most of the problems mentioned in this thread have to do with apps and VMs. I use a NAS for data storage, and I would not want to complicate it by adding apps or VMs. That's not the core mission of a NAS. I don't need a "jack of all trades" box. It seems safer to keep a NAS for NAS purposes and a separate box for other stuff, although admittedly that does boost the energy cost and carbon footprint a bit. In any event, Core has never been particularly good for running apps (containers) or VMs, so whatever problems Scale has with this are likely no better in Core.

Finally, I agree that it's inconvenient that rsync modules are being phased out. The claim that they're insecure is dubious given that they're no more insecure than NFS, which I suspect is the central use of most Scale installations.

That being said, switching rsync tasks to SSH is not difficult. See a separate post I made about this for steps to do it, based on my own experience. Moreover, I've found that there are some slight advantages with SSH compared to modules, having nothing to do with security. Namely, the file paths are stored on the Scale server rather than the remote host with SSH, which is better in my opinion because they're easy to back up through a configuration backup, whereas with modules, if the remote host system drive becomes unreadable, then the paths are lost unless one has a Clonezilla or other backup.

 

[winnielinnie]

The developers found it required a major change and there was risk involved.

This is what I mean. The risk to enterprise customers. This tentative approach (which makes sense for an enterprise clientele) is antithetical to rapid-development and introducing new features, as well as "let's just fix this bug already".

SCALE doesn't have this problem as much right now, since it's where most of the development is taking place.

It's Core that is being left behind. If there was a "Community" edition of Core, more "risks" could be taken. A "playground" where new features and bugs are fixed without the tentativeness of "risk to enterprise" customers. Enterprise can benefit from this down the line, as it becomes clear that a new feature or fixed bug is safe to merge into the enterprise codebase.

Right now, that's not the case. Instead it's:

"This is broken on Core."

"Fixed for SCALE. Too risky to fix for Core."



EDIT: But to the point of the post just above mine, the focus is on SCALE, so such a "Community" edition of Core (i.e, FreeBSD-based) is not part of the longterm vision.

 

[danb35]

Is there no workaround that is acceptable?

Sure, there's a workaround: don't use the shell through the GUI, and use SSH instead, which has been the advice of many on the forums, myself included, for many years. But why have you left such a thoroughly broken "feature" in place for over ten years? If you're not going to fix it, remove it--but keeping this brokenness in place for as long as it's been really reflects poorly on iX.

there was risk involved.

What risk? It's broken (to the point of being nearly unusable) already, and it's always been broken; it's hardly going to get more broken.

Edit: it could surely be claimed that this is a minor feature, and therefore that its brokenness is a minor problem. I don't think such a claim would be unrealistic--but that doesn't change the fact that the shell in the web UI has been nigh-unusable for over ten years. Nowhere during that time have you (plural, naturally, not you specifically) found it worth the effort to put in a working one. And it's not like you'd have to code your own; Proxmox (to name just one) has a web-based shell that works very well--control characters, copy/paste, etc., which are the major problems with the one in TrueNAS--and it's naturally F/OSS.

The "over ten years" thing really is important--for as long as iX has owned what was then FreeNAS, this "feature" has been broken; it's never been usable for anything more than the most trivial commands. This is an eternity in computer time. I don't care about the feature as such; I never use it--but that you care so little about a working UI that you've left this there this long really is a failure.

 

[morganL]

This is what I mean. The risk to enterprise customers. This tentative approach (which makes sense for an enterprise clientele) is antithetical to rapid-development and introducing new features, as well as "let's just fix this bug already".

SCALE doesn't have this problem as much right now, since it's where most of the development is taking place.

It's Core that is being left behind. If there was a "Community" edition of Core, more "risks" could be taken. A "playground" where new features and bugs are fixed without the tentativeness of "risk to enterprise" customers. Enterprise can benefit from this down the line, as it becomes clear that a new feature or fixed bug is safe to merge into the enterprise codebase.

Right now, that's not the case. Instead it's:

"This is broken on Core."

"Fixed for SCALE. Too risky to fix for Core."



EDIT: But to the point of the post just above mine, the focus is on SCALE, so such a "Community" edition of Core (i.e, FreeBSD-based) is not part of the longterm vision.

The focus is on SCALE for new features and capabilities....

The focus on CORE is maintaining existing capabilities and keeping them reliable and secure. We will backport improvements where feasible (e.g ZFS improvements).

I can see this issue (shell) issue is in a grey area... but I don't yet understand what is required to fix it. It may be code that we inherited that is not easily fixed. Would you agree with @danb35 that we should remove this shell option if we cannot fix?