RC1 is starting its final QA process

[morganL]

TrueNAS 13.0 RC1 has passed its code-freeze date and is going through its last QA phase this coming week.
There are over 100 bug fixes/improvements and only a couple of issues still being worked.

Expectation is that it will be available around April 12.

If you have a spare system for testing, the nightlies are available. https://download.freenas.org/13.0/MASTER/

The bug fixes are listed in Jira.

https://jira.ixsystems.com/browse/N...rsion = 13800 ORDER BY priority DESC, key ASC

 

[winnielinnie]

Is "CLI" considered stable for TrueNAS Core 13.0? Or is it still considered "experimental"?

 

[morganL]

The TrueNAS SCALE CLI is in SCALE only at this stage.

 

[winnielinnie]

The TrueNAS SCALE CLI is in SCALE only at this stage.

But it's already available even in TrueNAS Core 12.0-U8?

Code:

uname -a
FreeBSD truenas.nas 12.2-RELEASE-p12 FreeBSD 12.2-RELEASE-p12 ec84e0c52a1(HEAD) TRUENAS  amd64

cli

************************************************************
Software in ALPHA state, highly experimental.
No bugs/features being accepted at the moment.
************************************************************

truenas[]>  

 

[morganL]

The engineering post on the TrueNAS CLI is here: https://www.truenas.com/community/threads/introducing-the-new-truenas-cli.90741/

The 12.0 code may have had early stage development work, but it has not been completed. SCALE was the launch/test vehicle.

 

[kspare]

rc1 already? Ive been waiting for beta 2 to address the drive replacement button!

 

[Patrick M. Hausen]

@morganL Please make sure to get the latest FreeBSD patches in. That means at least FreeBSD-13.1-RELEASE or later. I know you are tracking -stable and will catch up, eventually. But there was a huge regression in epair that caused jails to go offline and would be a real nuisance to have in TrueNAS.

[Bug 261828] ifconfig: Remove new name display on interface name change

The fix is in the tree already and will definitely be part of -RELEASE.

Similarly there is a bug in bnxt that causes VLANs not to work at all unless the parent IF is explicitly configured "promisc". Most probably that will also make it into 13.1.

Kind regards,
Patrick

 

[Volts]

The epair fix landed in TrueNAS a couple weeks ago. I haven't experienced the forwarding problem again since.

if_epair: fix race condition on multi-core systems · truenas/os@1ddb6c5

As an unwanted side effect of the performance improvements in 24f0bfbad57b9, epair interfaces stop forwarding traffic on higher load levels when running on multi-core systems. This happens due to ...

github.com

 

[Louis2]

@morganL Please make sure to get the latest FreeBSD patches in. That means at least FreeBSD-13.1-RELEASE or later. I know you are tracking -stable and will catch up, eventually. But there was a huge regression in epair that caused jails to go offline and would be a real nuisance to have in TrueNAS.

[Bug 261828] ifconfig: Remove new name display on interface name change

The fix is in the tree already and will definitely be part of -RELEASE.

Similarly there is a bug in bnxt that causes VLANs not to work at all unless the parent IF is explicitly configured "promisc". Most probably that will also make it into 13.1.

Kind regards,
Patrick

Patrick, I would love to update my home NAS to TrueNas 13, however:
- that is not a test system
- Above I do read about a couple of serious issues
- You are talking about FreeBSD-13.1 release, which is to my knowledge released by the end of April, which is conflicting with releasing TrueNas 13 by 12 April which I did read somewhere ......
- lot of IPV6-issues as far as I know not fixed ... (and it seems that they will not be fixed in core ... )

Despite the last point, I would love to upgrade as soon as:
- the code is stable enough to assure the upgrade will not cause major problems, and surely will not lead to data loss
- it will be possible to upgrade to the stable version as soon as released.

What is your opinion on the actual status? Is it already safe to upgrade ????

 

[Patrick M. Hausen]

The TrueNAS-update on April 12th is RC1, not RELEASE. So I won't upgrade my productive home NAS running many jails immediately, but update my backup NAS, which is nothing but a ZFS replication target. And when TrueNAS 13.1 RELEASE hits the shelves, hopefully all these changes will be in there.

Honestly, it doesn't even take FreeBSD-13.1-RELEASE. If you pull stable/13 now, you already have the epair bugfix. What I don't know and did not check is how frequently iX pulls the FreeBSD source. So I hope @morganL can shed some light.

@Louis2 are you running jails and/or plugins with VNET? If you don't I guess it is safe to upgrade no matter if the fixes are in or not. If you have the Broadcom NetExtreme interface and VLANs, you already have the workaround in place - this is not a new bug. VLANs with bnxt never worked unless you explicitly configured "promisc".

So we have two different issues here - the bnxt thing has been in FreeBSD since the driver was introduced and I'm glad it finally got some attention. But people who have that hardware already have a workaround in place. The epair problem OTOH was introduced between FreeBSD 13.0 and 13.1 and I urge iXsystems to pull from upstream before TN 13.0 RELEASE if not before TN 13.0 RC1.

Edit: as @Volts wrote the epair fix is already in TrueNAS: #8

The bnxt issue (repeating myself) depends on a particular hardware and the use of VLANs and if you run that you already have the workaround in place.

HTH,
Patrick

 

[Louis2]

The TrueNAS-update on April 12th is RC1, not RELEASE. So I won't upgrade my productive home NAS running many jails immediately, but update my backup NAS, which is nothing but a ZFS replication target. And when TrueNAS 13.1 RELEASE hits the shelves, hopefully all these changes will be in there.

Honestly, it doesn't even take FreeBSD-13.1-RELEASE. If you pull stable/13 now, you already have the epair bugfix. What I don't know and did not check is how frequently iX pulls the FreeBSD source. So I hope @morganL can shed some light.

@Louis2 are you running jails and/or plugins with VNET? If you don't I guess it is safe to upgrade no matter if the fixes are in or not. If you have the Broadcom NetExtreme interface and VLANs, you already have the workaround in place - this is not a new bug. VLANs with bnxt never worked unless you explicitly configured "promisc".

So we have two different issues here - the bnxt thing has been in FreeBSD since the driver was introduced and I'm glad it finally got some attention. But people who have that hardware already have a workaround in place. The epair problem OTOH was introduced between FreeBSD 13.0 and 13.1 and I urge iXsystems to pull from upstream before TN 13.0 RELEASE if not before TN 13.0 RC1.

Edit: as @Volts wrote the epair fix is already in TrueNAS: #8

The bnxt issue (repeating myself) depends on a particular hardware and the use of VLANs and if you run that you already have the workaround in place.

HTH,
Patrick

Patrick,

My actual "active" NAS is only used as NAS without any jails or VM's. It is running 12U8. No real complains, however I do expect some improvements form V13 (performance and latest ZFS). Since I have more confidence in core than in scale, this NAS stays on core for at least another year. The more since it does provide all basic NAS-functionallity.

My second system which was and is intended as a combination of NAS and as server / host for "virtual machines". Is on pause at the moment. My idea to use TrueNAS for that failed.

Louis

 

[Volts]

Don't run alpha/beta/candidate/testing software on systems you can't afford to fiddle with. If you feel like asking "should I test this", the only safe answer anybody can give you is "no".

The nightlies are currently very stable for me, on a small home system with a mish-mash of jails. Performance is - anecdotally - great.

If you do test, it's easy to roll back to a previous boot environment.

If you do test, don't upgrade the features on your ZFS pools.

 

[morganL]

@morganL Please make sure to get the latest FreeBSD patches in. That means at least FreeBSD-13.1-RELEASE or later. I know you are tracking -stable and will catch up, eventually. But there was a huge regression in epair that caused jails to go offline and would be a real nuisance to have in TrueNAS.

[Bug 261828] ifconfig: Remove new name display on interface name change

The fix is in the tree already and will definitely be part of -RELEASE.

Similarly there is a bug in bnxt that causes VLANs not to work at all unless the parent IF is explicitly configured "promisc". Most probably that will also make it into 13.1.

Kind regards,
Patrick

Thanks for the heads-up.. please report the bugs if they don't get addressed. Out team is fairly pro-active on these issues and glad to see the epair issue is addressed.

As @Volts indicated, the general quality of TrueNAS 13.0 is pretty good because it is not a large deviation from TrueNAS 12.0. New OS issues will creep in and need to be resolved.

There have been a couple of recent security CVEs that have cropped up and we plan to address the RC1. These are expected to delay the RC1 until April 19th.

 

[Louis2]

OK,

My NAS is "off-cause" in the Green Zone, not exposed to the internet, so the CVE's might be less severe for me.

Given positive responses. I will install version 13 on my NAS. However since I will be away coming week, I will probably take that action in two weeks.

Not sure yet, if I will then install the beta, the nightly's or wait a few days for RC1

 

[Louis2]

I am playing around a bit with the "S3-service" on my 12U8 system, trying to store data on the NAS that way. Since for unknown reason the performance was "not great", I did try another S3-client (Duplicati 2).

Since I did not managed to get Duplicati working (issues with XML and perhaps "self signed" certificates), I decided to install 13B1 on a test system. Next to Duplicati I am testing with "S3 Browser" which in basic works.

I still do not have Duplicati working, however I did note that:
- minio (S3) was frequently stopping (crashing??) on 12U8, but not on 13B1
- and that where I had communication problems between "S3 Browser" and 12U8, they were gone with 13B1
- more issues to find and solve since I do not yet have Duplicati working

Given that result, I did manually upgrade 12U8 to 13B1.

Another issue is that as far as I can see, you can not select more than one IP-address in the S3-service (and also only one set of credentials), which is ^a bit^ limiting.

I am not going to do "testing" in at least the next two weeks, however if I notice strange things I will report.

 

[TooMuchData]

I am perplexed by 13.0-RC1. It is nine days late for release, and the single remaining issue has only recently been started. Meanwhile, the roadmap identifies 13.0 as a different version planned for release in less than three weeks. It is admittedly ancient history, but when I was in the software development game, release candidates were eventually improved into the actual release. So, what is going on? This is a question, not a complaint.

 

[hervon]

I am perplexed by 13.0-RC1. It is nine days late for release, and the single remaining issue has only recently been started. Meanwhile, the roadmap identifies 13.0 as a different version planned for release in less than three weeks. It is admittedly ancient history, but when I was in the software development game, release candidates were eventually improved into the actual release. So, what is going on? This is a question, not a complaint.

Answer 3 posts above:
"There have been a couple of recent security CVEs that have cropped up and we plan to address the RC1. These are expected to delay the RC1 until April 19th."

 

[TooMuchData]

But, RC1 is not really a release candidate since you have a separate 13.0 in the works. Am I missing something?

 

[Volts]

What are you referring to? What's separate?

 

[TooMuchData]

Look at the roadmap. There is an entry for 13.0-RC1 and another for 13.0.