zpool import hangs forever, help please!!!

[globus999]

You know FreeBSD is in Beta and not formal release plus it's free. You don't have to use it. Many people have informed you that your system is likely the cause of your problems and I think you need to move to UDF if you will be running 2GB RAM.

FreeBSD in beta? Where? www.freebsd.org clearly states "Production Release 8.2".

That old system is dead. However, for none of the reasons the posters suggested. I still don't know the reason but it was strictly within the mobo. Anyhow this is academic now. I did fork out a pretty penny for a brand new mobo/cpu/ram which *should* (emphasis on should) ran ZFS OK.

As for ZFS, there is nothing wrong with it, it was designed for the administrator level computer user, not the typical home user so the interface is purely command line like the old DOS days.

Not really, I refer you to the horse's mouth i.e. Sun where it was marketed as the ultimate FS that will serve anything from desktop to datacentre. Just check the post above. Don't want to repeat the link.

I thing the developers are doing a great job and eventually they will have a nice final product but until then it's Beta.

Ah! The light comes on! No, the "product" is not in beta. It has been released. And, it is a *product*. It is not an Open Source / Best Effort code anymore. Two standards completely different. It is *supposed* emphasis on *supposed* to be a commercial product. As a commercial product I expect and demand a much higher standard of quality. If you check the composition of the FreeNAS project you will notice that there are 16 people (presumably full time) in the project.

They even have a so-called Test Engineer.

Now, in a pure Open Source project, it being a community endeavor, I would be more than happy to take a completely different point of view. Why? simple, people are not trying to make a buck out of it. Consequently, is caveat emptor. If you can't stand the heat, get out of the kitchen.

However, this is most definitively not the case with FN8. FN8 was *specifically* created as a commercial product, for profit. Otherwise they would have used FN7.

I can tell you the following facts from 20 years worth of experience in the IT Quality Field (if you must know, I am also a Certified Software Quality Engineer among other degrees) if that helps you:

1 - This "product" has not been tested. I don't know what the Test Engineer is doing but it is not testing. If this person would have been in my team, it would have been fired long time ago. Critical functions and parameters were *not*, I repeat *not* tested.

2 - The documentation is poor to unacceptable. How do I know this? Because I have written more IT manuals and SOPs that I care to remember.

3 - The "official" support is non-existent. Not even for the people that are killing themselves and burning untold hours and $$$ to ensure that *they* get a better product. This is what's called and "externality" in economic terms (look it up). In other words, we are doing the work for them and they are not forking a cent.

I saw your trouble ticket 524 which slammed the developers for a unacceptable product.

Yes, I slammed them because it is the right thing to do. Again, it is a *product* not a hobby-open-source-diy thingy. In hindsight I am doing them a favor. Do you actually thing that any IT organization will touch a product like FN8 in its current state? None that I know of. They need to shape up and shape up quickly. Frankly, with 16 people in the team I can't even fathom what is that they are doing with their time.

I *wish* I had 16 people in any of the projects I am working on. That would be *really* nice. And, again, if you must know, the projects I work on are life-critical. If the software fails, people can die. So, you see, you can't get more critical than that.

After reading several of your postings I see you as a bitchy spoiled kid with nothing nice to say to anyone. My words of advice is if you really feel that way, move on to WHS 2011, it's a nice new product and I'm sure it works great, then you can contact MS when something doesn't work to your satisfaction.

Again, you are completely missing the point. I bring forward valid arguments, arguments that are basic, critical and well supported. Sure, perhaps my language is not as terse and politically correct as others, well, though! English is my fourth language and I gave up on getting it completely right long time ago.

And with regards to WHS I dislike MS products since I got burned too many times in the past. However, I now also dislike FN8 since it has been greatly oversold. Maybe 8.3 will be OK, maybe not.

Lastly, and for the record, I will cross-post this message in other forum since it is likely it will be nuked. This is, of course, beyond my control but a record should stay.

 

[joeschmuck]

Is this really not an open source project? Down the road are we looking at this project taking a turn and becoming a pay product? And in my opinion this is all still in Beta. I have yet to see a true release version, I think the "Release" version was pushed to gain popularity.

Not sure what you want to cross-post, doubt it will get nuked.

 

[matthewowen01]

FreeBSD in beta? Where? www.freebsd.org clearly states "Production Release 8.2".

That old system is dead. However, for none of the reasons the posters suggested. I still don't know the reason but it was strictly within the mobo. Anyhow this is academic now. I did fork out a pretty penny for a brand new mobo/cpu/ram which *should* (emphasis on should) ran ZFS OK.

FreeBSD 8.2 is not in beta, FreeNAS is, (or should be).

Not really, I refer you to the horse's mouth i.e. Sun where it was marketed as the ultimate FS that will serve anything from desktop to datacentre. Just check the post above. Don't want to repeat the link.

That's what they are working towards but zfs is not there yet. it's fantastic for a file server, home users will need to step up though. until you have (useless) pretty GUI's, you're not ready for home use.

Ah! The light comes on! No, the "product" is not in beta. It has been released. And, it is a *product*. It is not an Open Source / Best Effort code anymore. Two standards completely different. It is *supposed* emphasis on *supposed* to be a commercial product. As a commercial product I expect and demand a much higher standard of quality. If you check the composition of the FreeNAS project you will notice that there are 16 people (presumably full time) in the project.

They even have a so-called Test Engineer.

Now, in a pure Open Source project, it being a community endeavor, I would be more than happy to take a completely different point of view. Why? simple, people are not trying to make a buck out of it. Consequently, is caveat emptor. If you can't stand the heat, get out of the kitchen.

However, this is most definitively not the case with FN8. FN8 was *specifically* created as a commercial product, for profit. Otherwise they would have used FN7.

I can tell you the following facts from 20 years worth of experience in the IT Quality Field (if you must know, I am also a Certified Software Quality Engineer among other degrees) if that helps you:

1 - This "product" has not been tested. I don't know what the Test Engineer is doing but it is not testing. If this person would have been in my team, it would have been fired long time ago. Critical functions and parameters were *not*, I repeat *not* tested.

They are moving more commercial and using this as a testbed. Although i don't agree with your tact, i was disappointed that no one tested moving a drive to a different port for the first release. that should have been caught. it is what it is though.

2 - The documentation is poor to unacceptable. How do I know this? Because I have written more IT manuals and SOPs that I care to remember.

i too have found it a bit lacking.

3 - The "official" support is non-existent. Not even for the people that are killing themselves and burning untold hours and $$$ to ensure that *they* get a better product. This is what's called and "externality" in economic terms (look it up). In other words, we are doing the work for them and they are not forking a cent.

Yes, I slammed them because it is the right thing to do. Again, it is a *product* not a hobby-open-source-diy thingy. In hindsight I am doing them a favor. Do you actually thing that any IT organization will touch a product like FN8 in its current state? None that I know of. They need to shape up and shape up quickly. Frankly, with 16 people in the team I can't even fathom what is that they are doing with their time.

I *wish* I had 16 people in any of the projects I am working on. That would be *really* nice. And, again, if you must know, the projects I work on are life-critical. If the software fails, people can die. So, you see, you can't get more critical than that.

Again, you are completely missing the point. I bring forward valid arguments, arguments that are basic, critical and well supported. Sure, perhaps my language is not as terse and politically correct as others, well, though! English is my fourth language and I gave up on getting it completely right long time ago.

And with regards to WHS I dislike MS products since I got burned too many times in the past. However, I now also dislike FN8 since it has been greatly oversold. Maybe 8.3 will be OK, maybe not.

Lastly, and for the record, I will cross-post this message in other forum since it is likely it will be nuked. This is, of course, beyond my control but a record should stay.

4 languages, wow. i can ask for a piece of bread in french, if i want 2 i have to ask twice...

 

[joeschmuck]

Hey, the FreeBSD in beta was my typo, too much moving to/from FreeBSD and FreeNAS.

Hey, I like those useless GUIs! They make me feel all warm and fuzzy inside. Seriously, I don't mind cli provided there is a users manual for it. Back in the CPM and DOS days, everything have a good users manual. You didn't need to go buy extra books to figure something out and the internet was not around to ask a simple question. Give me a manual and I'll be happy, just not fuzzy inside.

 

[globus999]

Is this really not an open source project? Down the road are we looking at this project taking a turn and becoming a pay product? And in my opinion this is all still in Beta. I have yet to see a true release version, I think the "Release" version was pushed to gain popularity.

Not sure what you want to cross-post, doubt it will get nuked.

Let's clarify.

Formally speaking is a "product" based on Open Source software.... however, it's a product nevertheless. They are selling the Storage Appliance based on FN8, and, they are selling support. So, yes, it is a product. It just happens to have a different business model, not the traditional one.

Formally speaking it has been released. I fully agree that most likely it was released prematurely and with the purposes of making a profit / PR / getting to know it. Nothing wrong against an early release, however, the quality is just not there.

This is psichology 101: I, as a user and potential customer, don't see the problems. I only see features. I don't have experience in this particular NAS but I have a problem to solve. I don't have uudles of time or an army of techies to RFP (Request For Proposal - Short List - Evaluate - Try - etc.). I only see something that looks lean and mean and sufficiently basic so that a monkey could install it. Fantastic! Problem is it's not so. FN8 marketing method although not misleading is not truthful either.

Look, this is not rocket science, it is basic marketing.

FN is competing against other NAS products including MS.

How do yo do this? Simple, you differentiate. So, FN8 is being sold as light, effective, efficient and above all, free. Open Source does not really come into play for most IT shops at small to mid range enterprises. Nothing wrong with this approach.

Problem is, it's buggy and it will backfire. FN8 will get a bad rep and that will hit the bottom line.

The saddest part of this, is that it should not had to be this way. The correct SLC or SDLC (Software Life Cycle or Software Development Life Cycle) for a product like FN8 is the so-called Spiral (yes, you could do xtreme programming but there are other drawbacks).

The spiral dev lifecycle works quite well on new products. Heck! HP uses it.
The concept is simple. You build a basic ap. No frills. You make sure the quality is there. Then you sell it. Then you add features. Then you make sure the quality is there. You sell the upgrade..... and you keep going.

The key elements are these:

1 - Small number of features
2 - The quality is there (for which #1 is an enabler)

What did FN8 do?

They did a complete Web rework. Why? Flash in a pan? Too much effort for that. The number of developers working on the front end is about 40% vs 60% working on the back end. The percentages are all wrong. Most people using FN do not care about flash in a pan. They care about usability. Sure, the new web interface is somewhat usable, but, as an expert in usability I can tell you that it can be significantly improved. Users care about having a safe-harbor for their data. If the web interface looks a tad ugly (as long as it works) then that is not an issue.

They dumped CoreNAS (the to be FN based on Debian) and went with nanobsd. Why? Let's face it. Any Linux distro is much, much better supported than any BSD flavor. And what is nanobsd anyway? How many people use it? Where is the support outside a small number of geeks? Does not matter? But it does! Let me give you an example: the ataidle issue, it is correctly implemented in FreeBSD 8.2 but not in nanobsd. Check the Bug Tracker. Even the FN8 devs are stumped.

They could have gone with any embedded Linux distro which would have contained or at least would have been able to contain the latest ZFS code. They did not. They are tied to nanobsd. No latest ZFS code, no useful recovery tools.

They incorporated too many features into it. Too much complexity for a general re-write. Remember, spiral says keep it simple, make sure it works. For that, you need to have absolutely clear which features are vital and which ones can wait. And by wait, I don't mean 6 months, I mean 1 to 2 months turnover. What did they do? They mixed up nice-to-have features with critical ones and missed a quite a few vital ones.

Look, I don't deny any credit that is due to the devs. However (and I am saying this out of 20 years worth of experience), devs are NOT marketers and marketers are NOT users.

There is a vast universe between them and none of them speaks the same language.

Devs are geeks, they like to hack code. Back then, in the ol' days of punch cards and IBM 360 (look it up) we used to call them "code jockeys" and their product "spaghetti code". Times have changed, but human nature has not. A dev cares for code. Period.

Marketers care for sales. Sales are driven by:

1 - Flash in a pan (a door opener)
2 - Features (a sale closer)

They don't really understand users since marketers are not actually using the product day in and day out and they don't understand devs since they are not tech-inclined.

Users care for product solving. They don't really care for names, flashes or fancy features. They only care about solving problem X today. Tomorrow will be problem Y but that is tomorrow. Did you know, for example, that 96% of Excel functionality is never used? This means that roughly 96% of people that use Excel use only 4% of all functionality!

The people involved with FN should have fixated in that 4% first and made sure it works. They did not. Don't blame them tho... it's not easy. You have to be multidisciplinary with no ego involvement whatsoever, though call.

However, back to reality. FN8 in its current state is oversold and fixes are not forthcoming. How do you think a user that needs problem X solved today will react?

I mean, look at this statistic. Do you know what is the annoyance level for downtime for IT systems? Less tan 4%. In other words, if an IT system is down 4% of the time, you will have a lynch mob of users in your hand. If my experience and the experience of people that installed FN8 is to be accounted as a sample (sure, statistically speaking is flawed, but lets take a small leap of faith) then the downtime is about 50% or more!

And where does the downtime come from? Stability? No. Installation + Implementation. Once FN8 is stable it runs OK.... provided that you don't make any change (fat chance!).

But the user don't care! The user wants its problem X solved now!

That's the bottom line.

Now, informally speaking and with hindsight, deducing the level of testing FN8 probably went through (humm...) I would not even qualify it as Beta, I would label it as Alpha (i.e. not all critical features working / present).

But that's now. Going into it I had no way of knowing.

What's my objective? As a user, I want my problem X solved now.
How can I do this? By complaining loudly to the devs so that they can get their house in order and their plans straight asap, so that they can have a usable product which will benefit their bottom line. Win-Win, however, not painless.

Hope this provide some extra insight.

 

[globus999]

Hey, the FreeBSD in beta was my typo, too much moving to/from FreeBSD and FreeNAS.

Hey, I like those useless GUIs! They make me feel all warm and fuzzy inside. Seriously, I don't mind cli provided there is a users manual for it. Back in the CPM and DOS days, everything have a good users manual. You didn't need to go buy extra books to figure something out and the internet was not around to ask a simple question. Give me a manual and I'll be happy, just not fuzzy inside.

Yeah! sure! the GUI is nice. No complains from an eye candy point of view. I don't mind CLI either. Never bothered me. But you see you and I we are "power users" (don't laugh) there is a hierarchy there. We are a minority.

Most people are point and click. That's why a working GUI is absolutely critical (pls see my post below).

Anyhow, I don't mean to be rude, but sometimes one has to use stronger language to convey that something is vital.

Last example: ZFS tuning. The reality is that x32 does not work without tuning. For x64 is 50/50. Why on earth would somebody go to the trouble of FN8 and not implement ZFS? How could the devs miss this? Simple, they are devs.

I would bet you a box of doughnuts that they did not went to the trouble of minimally defining a set of Use Cases on which FN8 would be based. End result? Lacking problem.

However, I don't blame them. Paperwork is nobody's hobby, is that in something like this a modicum is... well... "strongly recommended"

 

[joeschmuck]

In my field paperwork is the bulk of the job and the product is supported by all that paperwork. I work in the tactical weapons field and when you're launching a weapon it had better function as expected. Sometimes there are errors in the code but that comes out during testing well before the product reaches the end user, that's one of the reasons why we test. Errors occur when you have literally hundreds of people involved in a product from several companies. We are allowed absolutely zero error for what product I work with, zero! And to date there has been zero errors as determined by follow on testing throughout the decades.

I know FreeNAS will not be held to the same standard as my company expects however it will be nice to see the final product with all bugs fixed.

Think I'm done with this thread, Cheers.

 

[pauldonovan]

And what is nanobsd anyway? How many people use it? Where is the support outside a small number of geeks? Does not matter? But it does! Let me give you an example: the ataidle issue, it is correctly implemented in FreeBSD 8.2 but not in nanobsd. Check the Bug Tracker. Even the FN8 devs are stumped.

They could have gone with any embedded Linux distro which would have contained or at least would have been able to contain the latest ZFS code. They did not. They are tied to nanobsd. No latest ZFS code, no useful recovery tools.

AFIAK, nanobsd isn't a different OS, it's just a particular way of building FreeBSD so that it takes up a small footprint. It's not a different code tree like OpenBSD. Any issues in nanobsd are due to the way the image has been created, not code issues. And the reason it doesn't have the latest ZFS is because neither does FreeBSD! There are no official releases of FreeBSD that have a version later than 15. The version 28 code was only considered stable a month or so ago, and won't be in a release until FreeBSD 9.0 in September. You can't blame the FreeNAS devs for that. FYI they have the v28 patch in the FreeNAS 8.1 branch already.

I've certainly had my frustrations with getting FreeNAS 8 working in a home environment, but after I've added a few packages I need and stopped the tweaking it's working OK for me now. I wish ZFSguru was still an active project, otherwise I'd still be using that.

Paul

 

[joeschmuck]

@ Paul

What extra packages did you add?

 

[pauldonovan]

@ Paul

What extra packages did you add?

bmon: for monitoring network performance. Has nice ascii graphs.
iftop: another network performance monitor
transmission-daemon: Bittorrent and nice web interface
mbmon: CPU and fan speed monitoring (my CPU doesn't work with coretemp variables in sysctl)
pytivo: not really a package, just a folder of python files. Serves video to my TiVo HD
ffmpeg: used by pytivo if it needs to do transcoding.

Paul

 

[Addrelyn]

Hi all, I would like to add something.
First, thanks for the solution to this problem. Maybe I will be able to get my data back...:D

I liked Freenas, the only problem is that for a home user like me, it needs a ridiculous amount of time just to achieve basic configuration, and if there's any problem... (and problem there are with this 8.0 version)
I will follow the advice given previously and go to WHS2011. My time is worth maybe 50$/hour. WHS will be quick to configure and costed me 50$, Freenas costed me already more than 50 hours I think...
No more hesitation for me!

Good luck to you all

PS: I'm not saying Freenas is bad, just that you need too much knowledge to use it properly. The total n00b that I am will now run back to Windows which is not perfect, far from it, but which work ok without fancy ZFS or anything.

 

[matthewowen01]

Hi all, I would like to add something.
First, thanks for the solution to this problem. Maybe I will be able to get my data back...:D

I liked Freenas, the only problem is that for a home user like me, it needs a ridiculous amount of time just to achieve basic configuration, and if there's any problem... (and problem there are with this 8.0 version)
I will follow the advice given previously and go to WHS2011. My time is worth maybe 50$/hour. WHS will be quick to configure and costed me 50$, Freenas costed me already more than 50 hours I think...
No more hesitation for me!

Good luck to you all

PS: I'm not saying Freenas is bad, just that you need too much knowledge to use it properly. The total n00b that I am will now run back to Windows which is not perfect, far from it, but which work ok without fancy ZFS or anything.

if you stop, you'll never learn. 3 years ago i abandoned windows as my server for, at the time Opensolaris. it was a painful process, i had little experience with any flavor of unix but i stuck with it. i'm not saying don't go with WHS, I'm saying make sure you have at least one unix box you use on a regular basis so you can get more comfortable with unix and hopefully one day not take 50 hours.

i agree the learning curve is very steep for FreeNAS 8. but once you get it down you're there. it took me about an hour to migrate from FreeNAS 7.2 to FreeNAS 8. you just need more practice.

good luck with WHS. drop back to let us know how it goes.

 

[Addrelyn]

Ok, for n00bs like me that don't understand anything (took me a long time to come up with this since I don't know linux, so installing zfs on a live cd...):

Burn a live cd of linux Ubuntu 11.04 (you can download it here)
http://ubuntu-fr.org/telechargement

Lunch your computer with this live cd. When asked, choose test linux instead of install linux

Open a terminal and type :

sudo -i
apt-add-repository ppa:dajhorn/zfs
apt-get update
apt-get install debootstrap ubuntu-zfs

When it's done, type:

zpool import -f <name of your pool>

That was it for me..
I'am now putting my datas on usb hard drives before destroying my pool and installing WHS.

Thank you all for your help!:D;)

 

[hrvoje@100posto.net]

What software can i use to recover files from zfs disk?
I had a FreeNAS 8 and my hardware gave out, when i reinstaled it and tried to import nothing happened.
Now i have all my data on that drive.
Can it be recovered using ubuntu and some software?
I have a backup but it's a month old
Please help

 

[joeschmuck]

You should be able to install FreeNAS 8 Beta 4 and AutoImport your ZFS volume without issue. Since you didn't specify which version of FreeNAS you are running I must assume you were using the "Release" version which is very old in terms of this project. Please install beta-4 or the latest nightly build.

 

[sonisame72]

Hi
I recently updated to FreeNas 8.01 beta4 (from beta1). I tried auto importing volumes, but I had no luck.

Using cli
when I do
zpool import -a -f
I get bunch one warnings saying
cannot mount '/DATA' : failed to create mountpoint
cannot mount 'DATA-2' : failed to create mountpoint
cannot mount 'DATA-3' : failed to create mountpoint

when I do
zfs list
I get
NAME USED AVAIL REFER MOUNTPOINT
DATA 1.51T 280G 1.51T /DATA
DATA-2 1.66T 129G 1.66T /DATA-2
DATA-3 388K 1.34TG 112K /DATA-3

I am not sure why I dont see /mnt/DATA-2 or something like that

Please help in mounting these individually mounted drives so that I can read my data again

 

[ProtoSD]

You need to do the import from the GUI or the mountpoint doesn't get created in the database and won't be created when you reboot. Use the 'Auto Import' feature in the GUI.

 

[sonisame72]

You need to do the import from the GUI or the mountpoint doesn't get created in the database and won't be created when you reboot. Use the 'Auto Import' feature in the GUI.

Thanks for replying, I just tried using auto importing function, but somehow I don't see any of the disks
I am attaching a screenshot of gui issue

 

[ProtoSD]

Strange, try exporting your volume from the command line before you import from the GUI.

 

[sonisame72]

Strange, try exporting your volume from the command line before you import from the GUI.

Thanks protosd
Export worked, but I got stuck at autoImport again, I am attaching the debug message as txt file. I am lost here in this large text file. Please help