Responding to your Feedback on 25.10, SMART, NVIDIA, and more | TrueNAS Tech Talk (T3) E045

So, any news on the SMB bug? It’s getting really annoying. Everyday I’m logging in and checking for updates, but still nothing.

I assume you mean the SMB2 query_directory with SMB_REOPEN causing jank?

We’re aware, we’ve isolated, we’ve hotfixed, and branched 25.10.0.1 - if you want the hotfix now, check https://ixsystems.atlassian.net/browse/NAS-138259 for the cherrypick and manual update instructions.

2 Likes

Yes, thanks, will try that if not too complicated to update manually. I’ve been using total commander as file manager for the last 25 years, and it happens much more than with Windows explorer, not sure why.

Edit: that was easy, just the package update. Rebooted, all seems fine now, thanks!

sorry what do u mean?

r u perhaps referring to this?

This week (Nov 18-21), there is a 25.10.0.1 hotfix release planned. It resolves about 10 of the more critical Goldeye bugs: SMB, Apps and general. It includes NAS-138259.”

i assumed smb will be fixed in a new release coming soon? if this is what u were referring to, then yes i don’t plan to install this goldeye till that new fix hits at least. even then i’m not sure if they resolved all the main issues, remains to be seen.

what is the default scheduling for the short and long smart tests running in background now, did they say?

If any SMART tests are running in the background, the interval between short tests is greater than eight days. A test system with a fresh install of 25.10.0 has been up almost 9 days, and the last SMART test logged by any of the spinners was 17 days’ of operation ago (at which time they were in a different system which had regularly-scheduled tests).

2 Likes

This feature change seems remarkably free of documentation nor has there been a coherent narrative as to the why.

All of which is ever more perplexing when we’re told that engineering came to this decision only after careful consideration, put together an elaborate transition plan, etc.

From a pure optics point of view, the most important thing seems to be not to admit that a change was perhaps wrong.

4 Likes

It’s “deniable plausibility” vs. “plausible deniability”.

1 Like

Nothing more frustrating than telling a bunch of self sufficient users that have happily adjusted those SMART knobs on their own for more than a decade “don’t worry about it, we got you covered with all this automatically in the background”. Truenas now has likely a significant number of users that have no idea what a smart test is which is probably why this is what’s happening. It’s a shame we can’t get the details of how it all works or someway to flip a switch to “take control”.

1 Like

I’m going to quibble with you on this one. I’d argue that the vast majority of paid and CE users enjoy TrueNAS not because it has the most polished UI, the most recent drivers, or even the latest support for whizbang graphics cards to run a home entertainment system. Other products like OMV allegedly do that better.

No, we’re here because of data integrity first and foremost. We don’t want to deal with bit rot, corruption, or data loss and we’re willing to make some compromises re: performance, driver support, and hardware support in return.

So of all the user groups out there, we’re likely among the most familiar with SMART tests, the interpretation thereof, and the use case for them.

As such, I’d be ok with smart tests being automatically scheduled for background execution as long as there was an advanced pane left in the GUI to fine tune the schedule. A even more useful enhancement would have been a better display and intelligent (AI-enhanced?) interpretation of SMART results, however.

1 Like

My point is they took away control and visibility which is not fun for our longer term users. I feel like scales maturity exploded the TrueNAS user base so we have a bunch more users that are following YouTube walkthroughs on how to put these systems together as an alternative to Qnap and Synology. I see the motivation to handle smart tests automatically to curb any unexpected failures for this “newer” crowd. I am guessing but I feel like the user base as a whole is less technically savvy than it was say 2 or 3 years ago. Again it’s the loss of control that is upsetting, not the move to automatically handle tests (good imo)

See you at Milliways.

1 Like

2 Likes

Please review the release notes and link to these in any discussion.

3 Likes

Ah, so we’re back to “the unmaintained app Scrutiny, which doesn’t do SMART self-tests at all, is the way to go.”

You just can’t stop digging, can you?

im no expert, but as a layman for my own particular use case,

i would use the UI in truenas to do a manual smart short or long, via the UI, since it’s easy to do.

thats all i ever used it for. doesnt seem like much, but when my NAS seagate hdd went bad recently, i needed to run a smart.

i’m assuming you can still run a manual smart at anytime but now only using commandline in shell yes? do u perhaps have to do it in tmux? or will regular shell work fine for that?

because in the previous method, when smart completes run, the task log when viewing smart in truenas will say that smart completed for short or long. thats how i would know it completed fine or not.

once that was verified, i would then use scrutiny to view the smart since that layout was more friendly for me to read.

as for smart scheduling, what worked for me was short smart weekly. long smart monthly.

and for my recently received brand new seagate ironwolf 12tb, i did the badblock scan via shell tmux. once done, i did the smart short then the smart long via truenas fangtooth UI. checked the smart in truenas ui, nothing mentioned. then checked scrutiny, also nothing bad.

Got my truenas up and working with data restored from backup, if anyone cares :blush:


:rofl:

1 Like

noticed this other thread for 25.10

hoping the smb fixes also solve this :thinking:

yes only have to update and it is fixed now

1 Like

Hey that’s great! I missed it totally since there doesn’t seem to be a link to this resource in the SMART GUI restoration request thread, it was not brought up in the second request thread either. A documentation link was not included with this podcast thread either.

Could you consider including links to documentation based on the content of various techtalks, if applicable? Especially if it’s a main topic of discussion… For example, if there is a Podcast whose main topic is the removal of the SMART GUI interface, it would be awesome to include a link to the documentation re: the SMART GUI removal.

Lastly, the documentation is full of useful information, though I suggest some further editing to clarify what the default intervals are that iXsystems has chosen, how the the user can check whether SMART tests have actually run, etc.

Team ixsystems is asking users to abandon a tried and true interface for a widely-used feature, I suggest they deserve more information re: how this new system is supposed to run and what issues (if any) to look for re: running SMART tests on a CRON job basis vs. whatever the default silent interval is.

1 Like

i think a few mentioned a newer nightly beta that is 25.11 or something? i assume u r referring to that? so good to know it was fixed ty.

:blush:

being a non beta tester i will still wait for the status to indicate the stability

right now i am on 25.04.2.5 CE it’s rock stable :saluting_face: