Will TrueNAS Core 13.1 have ZFS 2.2.2?

[victort]

Please keep me posted too.

I'm eyeing another product, but a fork of Core would be amazing.

FreeBSD + ZFS + jails = an excellent NAS solution
(Which is what made FreeNAS / TrueNAS Core so great.)

The stagnation of Core saddens me. Especially when simple, innocuous things don't get fixed under the rationale of "cannot risk breaking the software for our enterprise customers."

As far as new features that made it to SCALE, but not Core? Don't even get me started. I cannot accept the same excuse of "but our enterprise customers" when SCALE gets on/off toggles for SMB and NFS shares, but Core does not. Same with the "snapshot hold" feature in the GUI.

For instance, what if an enterprise customer requests on/off toggles for shares? Would iXsystems implement them? (Which they already did for SCALE, anyways.) What if an enterprise customer requests a checkbox to protect a snapshot with a "Hold"? Would iXsystems implement it? (Which, again, they already did for SCALE.)

Screenshot from another user's SCALE dashboard.

Other product?

 

[winnielinnie]

Suggest we start a thread and then make sure the bug tickets are in place.

I honestly had considered making a list, with descriptions, screenshots, and videos. But my experiences on the bug tracker have left a bad taste in my mouth, combined with the divergent futures of Core and SCALE.

It takes time and energy to plead your case, and feels deflating when they get brushed aside as "won't fix" because it's "too risky" or considered non-essential.

I might reply to the thread you linked. It depends. (I have a few "fixes" in mind.)

 

[morganL]

I honestly had considered making a list, with descriptions, screenshots, and videos. But my experiences on the bug tracker have left a bad taste in my mouth, combined with the divergent futures of Core and SCALE.

It takes time and energy to plead your case, and feels deflating when they get brushed aside as "won't fix" because it's "too risky" or considered non-essential.

I might reply to the thread you linked. It depends. (I have a few "fixes" in mind.)

iX Engineers have a tough job to do... they have to deliver software on time with quality and help our sales team get business. They don't want to be rude, but they have to make tough decisions about how to move forward efficiently. We don't have a fat and happy sponsor that is guaranteeing incomes and bonuses. The software is free and the systems being supported can be very complex.

Individual requests from non-customers get attention because there are many users that have a similar issue/requirement. The idea of the thread is to work out which issues are very popular amongst the community and identify how the fix can be provided most easily. Jira keeps us organized, but its probably not the best communication vehicle.

In any case, please be assured we want to help, but we do need to pay the bills. Looking forward to seeing your suggestion list.

 

[winnielinnie]

Looking forward to seeing your suggestion list.

Okay, before I create new ones for the thread, what about this one?

I included a video that depicts the exact issue:

Log in with Atlassian account

Log in to Jira, Confluence, and all other Atlassian Cloud products here. Not an Atlassian user? Sign up for free.

ixsystems.atlassian.net

I succinctly describe the issue.

It's clearly a bug / unexpected behavior.

I demonstrate it with a video.



Here is the video and description for those who cannot access the bug report, starting after this line.

___________________________​

Warning message for "unchecking" encryption does not revert if user selects "Cancel"​

When the user “unchecks” the box for encryption (when creating a new child dataset), they are greeted with a warning message that reads:

Warning
All data stored on this dataset will be decrypted and the dataset marked as non-encrypted. Do you want to continue?

However, if the user clicks “Cancel”, the checkbox will remain unselected.

This is not the expected behavior, and is obviously a bug.

Clicking cancel should revert the change.

Please see the attached video.

EDIT: I posted it in the other thread. I'll gauge from there if it's worth describing and recording other issues with TrueNAS Core / the web GUI.

 

[Davvo]

If there are minor fixes to do, now is the time to catalog them and agree on priority.

Will take some time these days to go fishing in the forum for a few posts that come to my mind about a few things that came up in the past few years. Expect the WebUI Shell to be there :P

 

[morganL]

Okay, before I create new ones for the thread, what about this one?

I included a video that depicts the exact issue:

Log in with Atlassian account

Log in to Jira, Confluence, and all other Atlassian Cloud products here. Not an Atlassian user? Sign up for free.

ixsystems.atlassian.net

I succinctly describe the issue.

It's clearly a bug / unexpected behavior.

I demonstrate it with a video.



Here is the video and description for those who cannot access the bug report, starting after this line.

___________________________​

Warning message for "unchecking" encryption does not revert if user selects "Cancel"​

When the user “unchecks” the box for encryption (when creating a new child dataset), they are greeted with a warning message that reads:


However, if the user clicks “Cancel”, the checkbox will remain unselected.

This is not the expected behavior, and is obviously a bug.

Clicking cancel should revert the change.

Please see the attached video.

View attachment 75520



EDIT: I posted it in the other thread. I'll gauge from there if it's worth describing and recording other issues with TrueNAS Core / the web GUI.

Its a great candidate for fixing.. can you paste this bug in that new thread for people to vote on. NAS-122461

I'd prefer we give the team a prioritized list and then we can see what we get done.

 

[winnielinnie]

Done. I edited my post in the other thread.

(However, the other bugs in Core / GUI do not yet have Jira tickets, since I'm not sure if it's worth it yet. I'd rather dump them in the thread with screenshots and videos first.)

You'll notice these aren't "feature" requests. They're bugs / unexpected behaviors associated with the GUI. (Another one yet to be posted, demonstrates the precarious nature of deleting snapshots in the GUI's Storage -> Snapshots page.)

 

[CJRoss]

Probably the Web GUI's shell, which is nothing short of catastrophic. The real tragedy is that it worked okay for a while way back in the early days of the then-new UI.

Yeah the WebUI shell. CORE has a few QoL improvements that could be fixed but aren't in the name of stability, and to me some look like minor changes that would hardly impact it. That's why I wanted to understand the point of enterprise users.

Can you give me a use case where you would use that instead of just ssh?

I've always avoided the web shell because it and my love of Ctrl-R command searching don't get along. :D

 

[Ericloewe]

I wouldn't, typically. Joe Clueless User panicking because he has a problem and was just told to get the output of a certain command? Happens frequently.

 

[Davvo]

Can you give me a use case where you would use that instead of just ssh?

I find time consuming explaining how to correctly post any output instead of a upside down screenshot or an incomplete and badly formatted walltext. Having to explain on top of that how to setup an ssh connection is sometimes a reason to hope someone with more energy notices the thread.

 

[Arwen]

Om regards to the GUI Shell "feature" - This is one reason why I think new users should automatically start to learn normal Unix command line. Not necessarily much in the beginning, perhaps simple zpool status and zfs list. And maybe local cp or mv for very large files.

But, it is my opinion that sometime during the life of their TrueNAS server, Unix command line via SSH will be REQUIRED. It may be as simple as copying and pasting a line from a forum user. (The GUI Shell "feature" does not support copy or paste if I remember correctly.) At present, ZFS & TrueNAS are not 100% GUI perfect.

While it may be that some users can go years, and potentially the entire life of their TrueNAS server without SSH access, we have seen too many users that NEED Unix command line access. Either to trouble shoot the problem, or fix the problem. And some of those users persist in misspelling the commands, (because they did not / could not copy & paste), or not understanding what to do.


When I researched my first NAS back around 2005, I was disappointed with most of the options for home users. I ended up with an Infrant ReadyNAS 1000S, (later bought by Netgear). It supported NFS AND SSH logins, as well as "root" on the NAS. When it was time to replace that, I was pleased with FreeNAS and FreeNAS Mini features, including SSH in and ZFS boot, (in 9.3 or was that 9.10?).

 

[Ericloewe]

It's a tricky thing, between fear and legitimate risk of foot-shooting, but I do agree that being exposed to the Unix shell is a good thing. I guess it helps that ZFS' tools don't make me want to rip my hair out (*cough* mdadm).

 

[Davvo]

But, it is my opinion that sometime during the life of their TrueNAS server, Unix command line via SSH will be REQUIRED.

I agree, but I don't believe that panicked users who are experiencing a (possibly critical) issue are in the ideal conditions for doing so. The incompleteness of TN's GUI is imho the main cause of other products success: a polished UI helps you greatly with the majority of home users. Enterprise people (iX's customer target... I believe) might not care, and if iX doesn't care as much for average home users (for a number of possible reasons, ie limited resources) I can't blame them... but I still think it's a bit of a wasted opportunity.

Again, I want to make it clear that I am not criticizing or blaming iX: they have a model of business, which as far as I understand has proven to work, and I totally respect that. People can't live on passion alone, and engineers are still people... right?​

 

[Etorix]

To put it another way: Even if a SSH session is better, the web shell is there, conveniently at hand.
It's fine for casually typing a lone command (zpool status, smartctl -a,…).
But it would help a lot if the casual user could copy clean output from the web shell to report on his issue rather than having to learn extra tricks while panicking whether he'll ever get his data back!

 

[Arwen]

Exactly, this is why;

This is one reason why I think new users should automatically start to learn normal Unix command line. Not necessarily much in the beginning, perhaps simple zpool status and zfs list. And maybe local cp or mv for very large files.

Then when something bad happens, they can both understand how to log in via SSH and supply any requested information.

Basically, my opinion is that if they don't want to learn minimal Unix, ever, then TrueNAS, SCALE or Core, is a poor choice for SOHO NAS. This might change as more GUI improvements are made. Yet I think it will take more than a year or 2 for that to happen, (if possible).

 

[winnielinnie]

1. Go to System -> Advanced
2. Check "Show Console Messages"
3. Save
4. Click on the bottom of the screen where there is a small buffer of console messages.

Notice how clean the font is?

Notice it spans across the entire width of your page? (It doesn't "wrap" at 1/4th of your screen.)

Notice you can accurately copy text?

Not sure why the "Shell" is a broken feature, yet they clearly have the basics already working in their "Console Messages".

 

[CJRoss]

I was only thinking of normal user usage of TrueNAS and hadn't considered helping a new user who is having problems. That definitely makes sense.


Speaking of console messages, I've noticed that it's rather hit or miss whether something will get logged to syslog. I know one of the services has an option to include it's logs in syslog (SMB, I think) but most of the others don't that I can see.

 

[Patrick M. Hausen]

Without SSH how are you ever going to use jails? You need to edit configuration files inside the jails from the host before you can even SSH into the jail and continue from there. I always feel that creeping dread when I catch a novice user asking how to edit e.g. Nextcloud's config.php via SMB access from their Windows PC. Recipe for disaster.^M

 

[Ericloewe]

I always feel that creeping dread

Good description.

 

[Apollo]

Without SSH how are you ever going to use jails? You need to edit configuration files inside the jails from the host before you can even SSH into the jail and continue from there. I always feel that creeping dread when I catch a novice user asking how to edit e.g. Nextcloud's config.php via SMB access from their Windows PC. Recipe for disaster.^M

You can get away by using "mc" ("Midnight Commander"), usually over SSH, or within Web GUI Shell and go to the jails relevant location within the "iocage" top level location.