ZFS: can't read MOS of pool foo

[aufalien]

Perhaps if you post details about your NFS problems, someone more knowledgeable can help out.

That's just it, the problem is as of yet, unknown. The screen shot from earlier shows the error and since this thread really only has you and I on it, I doubt any one has seen this before or know what it means.

I've replicated large +120TB data sets before via nc on FreeNAS so it's not an issue doing it quickly. My question was simply about replicating.

I'm fairly sure that I cannot replicate a v12 dataset to a v11 dataset due to feature flags. Or am I wrong?

But can I replicate either a v12 pool to a v11 pool or a v12 pool to a v11 dataset.

Let me know your thoughts on these scenarios.

 

[Spearfoot]

That's just it, the problem is as of yet, unknown. The screen shot from earlier shows the error and since this thread really only has you and I on it, I doubt any one has seen this before or know what it means.

I've replicated large +120TB data sets before via nc on FreeNAS so it's not an issue doing it quickly. My question was simply about replicating.

But can I replicate either a v12 pool to a v11 pool or a v12 pool to a v11 dataset.

Let me know your thoughts on these scenarios.

I don't know, as I've never tried to do that before.

 

[Spearfoot]

I've asked around a little, and been advised to suggest opening a ticket: a 'trampoline' panic and random reboots are bad JuJu!

But here's what I can do: I have a smallish pool on a 'sandbox' 12.0-U5.1 server; I will promote it to the latest feature set. Then I will try replicating one of its datasets back to one of my 11.2-U8 servers -- and see what happens! If it works, then you will presumably have a pathway to revert to 11.2-U8

 

[aufalien]

I've asked around a little, and been advised to suggest opening a ticket: a 'trampoline' panic and random reboots are bad JuJu!

But here's what I can do: I have a smallish pool on a 'sandbox' 12.0-U5.1 server; I will promote it to the latest feature set. Then I will try replicating one of its datasets back to one of my 11.2-U8 servers -- and see what happens! If it works, then you will presumably have a pathway to revert to 11.2-U8

Why thank you!

I'd rather not put you out as I got word that I can indeed replicate newer pool to older pool which makes total sense.

The ticket submissions have never gotten me anywhere so I'll avoid it. I've had a few over the years and have only heard crickets chirping as a result.

 

[Spearfoot]

Can't say I blame you...

Regarding confirmation: The FreeBSD Manual Pages, here:

zpool-features

www.freebsd.org

Say that:

The pool format does not affect file system version compatibility or the ability to send file systems between pools.

So I think you're "Good to Go".

 

[aufalien]

Can't say I blame you...

Regarding confirmation: The FreeBSD Manual Pages, here:

zpool-features

www.freebsd.org

Say that:So I think you're "Good to Go".

Well, all in all I am not good to go.
This error is fairly a large one as I've a 130TB server to downgrade to an 11 release.
I'm starting to see this behavior in my other large 12.x TrueNAS servers which is a nightmare.
So not good to go.

 

[Spearfoot]

Well, all in all I am not good to go.
This error is fairly a large one as I've a 130TB server to downgrade to an 11 release.
I'm starting to see this behavior in my other large 12.x TrueNAS servers which is a nightmare.
So not good to go.

Poor choice of words on my part... I only meant that you have a way forward out of the quagmire you're stuck in now.

 

[aufalien]

Poor choice of words on my part... I only meant that you have a way forward out of the quagmire you're stuck in now.

No worries man. You are the ONLY one really caring enough to reply to my thread and I am thankful!
I'm just stressed over this. Shoulda gone FreeBSD or Linux all those years ago as a UI is over rated.

 

[aufalien]

Interestingly, my other major servers are running 12.0-U4 and have been up and rocking for several months. I'm going to do a fresh install of 12.0-U4 and import the pool. At that point I'll see how it behaves. However I've a long SMART scheduled for the evening and will see if that by chance catches anything. I run em monthly w/o issue but one never knows.

If my tests with 12.0-U4 fail, I'll also try 11.2-U8 knowing my current datasets will be read only. However I should be able to create another dataset and copy from the 12 datatset to the 11 dataset as I've enough local space to do a few datasets at a time while testing.

By the way, you mentioned testing so would you be able to see if one can replicate from a 12 dataset to an 11 dataset? I'm hoping that the unsupported feature flags would be stripped out etc... Or perhaps a 12 dataset to an 11 pool?

 

[Spearfoot]

By the way, you mentioned testing so would you be able to see if one can replicate from a 12 dataset to an 11 dataset? I'm hoping that the unsupported feature flags would be stripped out etc... Or perhaps a 12 dataset to an 11 pool?

As I understand it, pool features aren't involved when replicating datasets.

I successfully ran the test just now, replicating dataset tank/boomer_nfs from my TrueNAS 12.0-U5.1 server BOOMER to dataset dozer/boomer_nfs on my FreeNAS 11.2-U8 server BACON. I first deleted and then re-created the dataset on BACON so that it started out empty. My script:

Code:

logfile=./repl-to-dozer.log
zfs snapshot -r tank/boomer_nfs@backsnap_boomer_nfs | tee -a ${logfile}
zfs send -v -R tank/boomer_nfs@backsnap_boomer_nfs | pv | ssh bacon zfs receive -v -F dozer/boomer_nfs | tee -a ${logfile}
zfs destroy tank/boomer_nfs@backsnap_boomer_nfs | tee -a ${logfile}

And the log output:

Code:

receiving full stream of tank/boomer_nfs@auto-20210925.1412-4w into dozer/boomer_nfs@auto-20210925.1412-4w
received 42.2GB stream in 207 seconds (209MB/sec)
receiving incremental stream of tank/boomer_nfs@auto-20210926.1412-4w into dozer/boomer_nfs@auto-20210926.1412-4w
received 853MB stream in 6 seconds (142MB/sec)
receiving incremental stream of tank/boomer_nfs@auto-20210928.0900-4w into dozer/boomer_nfs@auto-20210928.0900-4w
received 502MB stream in 4 seconds (125MB/sec)
receiving incremental stream of tank/boomer_nfs@auto-20210929.0900-4w into dozer/boomer_nfs@auto-20210929.0900-4w
received 1.03GB stream in 5 seconds (210MB/sec)
receiving incremental stream of tank/boomer_nfs@auto-20210930.0900-4w into dozer/boomer_nfs@auto-20210930.0900-4w
received 420MB stream in 4 seconds (105MB/sec)
receiving incremental stream of tank/boomer_nfs@auto-20211001.0900-4w into dozer/boomer_nfs@auto-20211001.0900-4w
received 99.5MB stream in 1 seconds (99.5MB/sec)
receiving incremental stream of tank/boomer_nfs@auto-20211002.0900-4w into dozer/boomer_nfs@auto-20211002.0900-4w
received 234MB stream in 2 seconds (117MB/sec)
receiving incremental stream of tank/boomer_nfs@auto-20211003.0900-4w into dozer/boomer_nfs@auto-20211003.0900-4w
received 328MB stream in 3 seconds (109MB/sec)
receiving incremental stream of tank/boomer_nfs@backsnap_boomer_nfs into dozer/boomer_nfs@backsnap_boomer_nfs
received 321MB stream in 2 seconds (161MB/sec)

The pool features on pool dozer remained unchanged.

So... if you can keep your 12.0-U5.1 system up and running long enough, you should be able to replicate its datasets to another system or pool.

 

[aufalien]

Thank you, this is very helpful!

I was curious, since my pool no longer shows the error that I had initially opened this thread on, I'd like to close it.

However I'd also like to start a new thread showing a Quicktims of my boot up with some very odd errors in hopes that some one may be able to help tell me what they mean.

Is it possible to post a 35MB Quicktime in this forum?

 

[Spearfoot]

Thank you, this is very helpful!

I was curious, since my pool no longer shows the error that I had initially opened this thread on, I'd like to close it.

However I'd also like to start a new thread showing a Quicktims of my boot up with some very odd errors in hopes that some one may be able to help tell me what they mean.

Is it possible to post a 35MB Quicktime in this forum?

I don't know what the file size limit is, but I suspect 35MB may be too large. You can always just give it a try and see what happens...

If it's too large for the forum software, then I suppose you could try linking to it at a public filesharing service.

 

[jgreco]

What's even more abysmal is that seemingly no one here has ever brought this up with IX Systems who can then speak with the upstream provider.

Just for your info, none of us here have any sort of magic backchannel to iXsystems. If there is not an "iXsystems" badge under a name, then we basically have the same recourse as you do, which is to file Jira tickets or to try to catch the attention of someone who IS an iX staffer. Frustrating for us too.

 

[aufalien]

Just for your info, none of us here have any sort of magic backchannel to iXsystems. If there is not an "iXsystems" badge under a name, then we basically have the same recourse as you do, which is to file Jira tickets or to try to catch the attention of someone who IS an iX staffer. Frustrating for us too.

Yes I know. Hoping for the latter. But I do understand what you are saying. Besides, I'm inclined to go with the community as I've spoken with IXSystems for quotes and info. And honestly was a bit disappointed by whom I was speaking with in terms of there knowledge and recommendations.

I for one have seen some VERY VERY VERY sharp and informed ppl here on the public forums which is why I go with open source in general. You get a broader and more knowledgeable base than from a company.

I've seen some really good posts from you over the years in fact.

 

[jgreco]

I've seen some really good posts from you over the years in fact.

Well, I sure hope so. That's mainly why I hang around here. ;-)