Any thoughts on moving Beta 3 up a few days?

[diskdiddler]

https://redmine.ixsystems.com/issues?fixed_version_id=570&set_filter=1&status_id=o
Only 5 left to go. It's been under 10 for maybe a week.

That 'crash / reboot entire system' when deleting a large amount of files bug (fixed in Beta 3), isn't a particularly nice one to have. I know it's a beta but we still have RC1 to go anyhow.

 

[Tim1962]

I know the correct, sensible and "I agree with " answer is release when safe and sensible and ready etc.

However I also agree that the residual issues seem relatively minor compared to the issues that I am waiting for resolution in Beta3, namely as diskdiddler identified the deleting file bug and the failure of the Nextcloud plugin for looping pages

WAAAAY above my paygrade but a beta3 and beta3b would be lovely...

 

[dlavigne]

There are still 37 tickets in testing so it is not likely it will be before next Monday. This is a short week and we don't like to release late in the week.

 

[Tim1962]

I didn't really expect a response (but thanks) and I am sure you are aware some bugs are more buggy than others. I'll let you get back to the real job

 

[diskdiddler]

There are still 37 tickets in testing so it is not likely it will be before next Monday. This is a short week and we don't like to release late in the week.

I feel pretty dumb.
Where is that visible? I see only 3 open tickets for it?
Surely if it's in testing, it should still be open?

 

[kdragon75]

Please don't ever rush a release. in fact, I would prefer they didn't have deadlines. Release when you feel it's ready. Like pfSense.
That said, I'm getting anxious for 2.4.4 with routed IPSec! Even then I'll be waiting for p1 (u1 in FreeNAS parlance)

 

[diskdiddler]

Please don't ever rush a release. in fact, I would prefer they didn't have deadlines. Release when you feel it's ready. Like pfSense.
That said, I'm getting anxious for 2.4.4 with routed IPSec! Even then I'll be waiting for p1 (u1 in FreeNAS parlance)

Normally I would agree with this, but now there's X amount of users who are on a beta which could crash at any moment if they delete a folder full of files.
Also, they're done heaps of work since B2 to B3 and closed many tickets, there's very, very few left, of which none seem particularly important.

Don't rush the RC1, don't rush the final 11.2 - but pushing something ahead a few days might be worth it. Honestly I'm not sure why so many days remain for B3 as it is.

 

[kdragon75]

on a beta

That's what happens with beta...

 

[hervon]

Normally I would agree with this, but now there's X amount of users who are on a beta which could crash at any moment if they delete a folder full of files.
Also, they're done heaps of work since B2 to B3 and closed many tickets, there's very, very few left, of which none seem particularly important.

Don't rush the RC1, don't rush the final 11.2 - but pushing something ahead a few days might be worth it. Honestly I'm not sure why so many days remain for B3 as it is.

Go for nighlies ?

 

[diskdiddler]

That's what happens with beta...

Yes I realise that, this is a single exception, again, because I saw only 5 tickets in the queue and it's been under 10 for over a week already with another week to go.
The beta is now becoming a liability to test, due to the fault in it.
I'm not demanding a release, I'm suggesting a release. Alternatively make it go to beta 4, I don't particularly care if there's 11 of them. However we have a pretty nasty bug, in beta 2, that's been there for weeks, has been fixed for weeks, but it's being held back.

 

[rungekutta]

That's what happens with beta...

An easy answer, but... it’s a NAS. Not a fancy hypervisor, or complex container to accommodate generic services. Perhaps that’s where it’s going but... can’t help in the meanwhile to be a little surprised to see that kind of bug even in a public beta. At least it begs the question of the robustness and coverage of automated testing, including smoke/load testing with a few different scenarios of clients. That would reasonably have caught the 11.1-U6 issue also of erratic performance with Mac clients when under load.

Again easy to argue from the sidelines... but I would have thought for a NAS you want to beat the cr-p out of every version through automated load testing, and verify behaviour, before anything is released, including public betas...

 

[kdragon75]

An easy answer, but... it’s a NAS. Not a fancy hypervisor, or complex container to accommodate generic services. Perhaps that’s where it’s going but... can’t help in the meanwhile to be a little surprised to see that kind of bug even in a public beta. At least it begs the question of the robustness and coverage of automated testing, including smoke/load testing with a few different scenarios of clients. That would reasonably have caught the 11.1-U6 issue also of erratic performance with Mac clients when under load.

Again easy to argue from the sidelines... but I would have thought for a NAS you want to beat the cr-p out of every version through automated load testing, and verify behaviour, before anything is released, including public betas...

I don't disagree with any of that. But the point I'm making is that if you trust you mission critical data or ANY data you care about even from the perspective of availability, you have NO business running beta software. It's that simple. I'm still on 11.1u4 and will be until at LEAST 11.2u1.

 

[rungekutta]

I don't disagree with any of that. But the point I'm making is that if you trust you mission critical data or ANY data you care about even from the perspective of availability, you have NO business running beta software. It's that simple. I'm still on 11.1u4 and will be until at LEAST 11.2u1.

Agreed and also supported by empirical evidence... I’ll await the U1 if not U2 as well before jumping on next major version.

 

[rungekutta]

PS @kdragon75 why does 11.1-U5 suck for you (as per your sig)? Didn’t U4 have nasty L2ARC infinite growth weirdness?

 

[danb35]

That's what happens with beta

True enough. But the counter is that the beta is what iX is pushing on the download page, falsely suggesting that you can't do VMs, jails, or plugins with 11.1.

 

[kdragon75]

PS @kdragon75 why does 11.1-U5 suck for you (as per your sig)? Didn’t U4 have nasty L2ARC infinite growth weirdness?

I don't use a L2ARC.

 

[Tim1962]

Just FYI and the original post...

Gritted teeth
Backup up settings
Cloned 11.2-beta2 environment
Moved server to 11. nightlies
Deleted my offending sparsebundle with millions (?) of bands
Rebooted back into 11.2-beta2

All sorted
PS used rm -rfv (v for verbose otherwise the screen just hangs there for a long time!!!)

Unfortunately (for me) the nextcloud plugin still fails with an infinite loop in the nightlies train

HTH Tim

 

[diskdiddler]

I don't disagree with any of that. But the point I'm making is that if you trust you mission critical data or ANY data you care about even from the perspective of availability, you have NO business running beta software. It's that simple. I'm still on 11.1u4 and will be until at LEAST 11.2u1.

So you've made it clear from your post, that you're not someone into testing betas. So I have to ask, what business do you have really, telling us,....the beta testers, that we can't ask for a new BETA?
I'm not (in any way!) asking or demanding for a faster release to the final version. AT ALL.
I'm wanting us, to have a better version to keep testing, for folks like you.
If anything I've posted multiple times here "please, please don't rush things, make sure stuff is all completed and tested" but that's for final release, not beta.


We're not stupid, we know betas have risks.
Infact, we've now encountered a fairly reasonable risk in beta 2
That risk has been fixed, for quite a bit now.
It would be nice, to move past that risk and move to a new beta. After all Beta 2 fixed over 200 bugs if I recall, surely that's worthy of testing.

This thread is quickly becoming irrelevant, as we're down to 6 days left, but I'd still like to see it bumped up.

 

[danb35]

So you've made it clear from your post, that you're not someone into testing betas.

No, that's not remotely what he said. He said (correctly) that you shouldn't use a beta in production. If a software bug (up to and including a pool-destroying bug) is going to give you heartburn, don't use a beta. Your stated argument for moving up the target for B3 is that there are lots of people on B2 in trouble with a bug--and the answer to that is that if the bug is a problem for them, they shouldn't be using Banything.

iX has been giving all kinds of mixed signals with this beta, more so than with previous releases--but it's still a beta, and clearly marked as such.

 

[diskdiddler]

No, that's not remotely <SNIP> ked as such.

You people _continue_ to not get it.
What possible difference does it make, if a new BETA is shipped earlier.

They can do 1 beta, 2 betas, 194,312 betas, they are beta. That's the point of a beta, for testing.
There's now a bug which is prohibiting extensive testing which could be easily solved by /rolling out another beta/

People in this thread need to stop deflecting at "told you so" attitudes, when we're only asking for more /betas/ not rushed final software.