Another day, another samba ... 3.2.11

[soorploom]

With bated breath, I noticed the updates. Could it be that z3.2 will now work as a file server, actually being able to serve...files?

After all, we're now into the third month since z3.2 was released and now with Zentyal's version of samba at number eleven, surely, at last, it will actually work. After all, a server without file sharing capability is like a pig without trotters.

Anyway, updates all done, even rebooted it which was quite brave as the system is remote, and keyed in the details of a small testshare under /mnt. What happens? Well, back to,

Some modules reported error when saving changes . More information on the logs in /var/log/zentyal/
The following modules failed while saving their changes, their state is unknown: samba

I'm not even going to post log output as zentyal.log has told me more made up stories and lies than a politician caught with his hand in the till. As far as posting another bug, pointless, as samba entries have moved in an taken that over.

Still doesn't work. samba 3.2.99 anyone?

Enjoy your day.

[lhx1]

I have to agree. This is a fairly central part of an operating system that is supposed to be ready for a production environment. It clearly isn't.

I feel zentyal have really dropped the ball in the 3.2 release. Perhaps they are rushing too hard to release software that should still be in beta. Maybe even alpha.

The current release just isn't up to snuff and I am sincerely regretting upgrading my system.

[christian]

I do understand your point but it made me raise my eyebrows.

Let's be fair. Integrating something like Samba 4 which has much more constraints than previous version is a real challenge, given the fact that  Zentyal has its own constraints too.

I'm not saying this as an excuse however 

What is somewhat interesting is that lot of people want to have the very last version of everything, stable from the beginning.
When Zentyal decided to rely on Samba 4, it was already clear, at least to me, some month ago, that this was a real challenge. See my previous comments on this topic.

In my understanding, one of the reason why Zentyal is rushing toward news features is that customers or potential customers are asking for it.
You are currently asking for file server able to share file (which is perfectly fine, of course) but when Zentyal was running (stable) Samba 3, the main request was "we can't survive without true Windows domain emulation, thus (beta) Samba 4)"

In addition, I'm pretty sure some 3.2 users are happy with their current implementation, meaning you are facing problems with your deployment but some are running 3.2 file sharing service.

I fully share that 3.2 being released, it is of course expected to work smoothly enough to be used in prod  but given the amount of problem reported in this forum around 3.2, wise decision is perhaps not to run it like this, at least for the time being, or to buy support from Zentyal if you need assurance that it will not interrupt services you are running.

I'm a quite happy Zentyal user, running 2.2 and my file server is even not Zentyal 

[Lonniebiz]

Zentyal 3.2 file sharing and permissions have been a disaster, but at least I can see they are actively working on it. I see frequent updates for these components.

One thing's for sure, I won't be quick to hit the upgrade button when future releases come out. 3.0 was very decent, and I wish I would have just stayed on it until all these issues were completely worked out.

The main thing I hope Zentyal learns from 3.2, is NOT to call something a release until they've confirmed it works. 3.2 should still be labeled beta (right now).

[soorploom]

Christian,

Your points are noted. You say, regarding the total lack of workable sharing in z3.2, "Let's be fair. Integrating something like Samba 4 which has much more constraints than previous version is a real challenge, given the fact that  Zentyal has its own constraints too. I'm not saying this as an excuse however".

You then continue to make excuses in Zentyal's favour.

No doubt it was and is a real challenge to come up with a reasonable product which reflects Microsoft's offerings but my whole point is that after numerous patches over a period of more than two months, file sharing, the essence of a server, does not work. At best, z3.2 is fragile at worst, unusuable and I'm afraid to say that mostly, it is the latter.

Now, had it been released as a beta, a test, have a bash and find some bugs thing, well yes, no doubt some would have given it a go but it wasn't. It was released as a fully functioning, well upgraded suite of applications which addressed many of the shortcomings of the previous versions. In Zentyal's own advertising splurge for z3.2, "Are you looking for an IT infrastructure that simply works".

It doesn't work, simply or otherwise.

Now, had there been some information from Zentyal as to how they can see the problems being addressed and a bit of an idea as to when, that would be better, much better but thus far, nothing, zilch, silence.

The logs cannot be trusted to offer some reasonable cause for failure, zentyal.log seems to have a mind of its own and produces some fantastic output which differs depending on which way the wind blows.

I am happy to give z3.2 a bit longer to settle but in the interim, some feedback from Zentyal would be useful. Other than that, I will have no other option but to abandon z3.2.

[Cellobita]

I'm not in the habit of posting "me too" messages, but I absolutely agree with your assessment; the older versions of Zentyal had less functionality, but the one it had did work. I am not a programmer, and my Linux/Samba knowledge is average at best, but it seems unbelievable that they could release 3.2 without this very basic functionality working.

At the very least there should be some sort of official recognition, by the developers, of the most pressing issues with the software, and a timetable for fixing them - Zentyal has an excellent window of opportunity (about five months, until next April) to make inroads on the large customer base that Microsoft abandoned when they dropped their SBS product line, but the basics have to work, consistently and reliably: can't sell the whole Linux replacement idea to them otherwise.

Just my 2 cents...

Marcello
Sao Paulo - SP - Brazil
   

[christian]

My point is not to find excuses    but to say, although I do understand your frustration, that it had to be expected from the beginning when Zentyal announced they were going to release next version with "Samba beta" inside. In the meantime, Samba changed quite a lot therefore Zentyal has to:
- deal with Samba bugs and fixes
- adapt their own code to changes at Samba level

These are not excuses but facts.

I may be wrong but the way I perceive Zentyal is that it brings an additional layer on top of open source components. No more nor less than this. This is supposed to bring some consistency and ease global administration, drawback being lack of flexibility.
With such vision, it helps (at least me) to select the right version and decide how to implement features I need.
- Zentyal file server is not what I'm looking for, not flexible enough and furthermore restricted to CIFS. I'm running another file server relying on Zentyal LDAP.
- I don't need Windows domain stuff: Zentyal 2.2 with Samba 3 fits my needs.
- When I need to look at logs, I'm not reading Zentyal logs first but the one generated by components under investigation.

This is what I try to explain, adding that Zentyal is rushing toward more "advanced" features because potential customers are asking for it but on the other hand, they do not understand that this is more complex, risky and potentially unstable.
In order to deal with this extra complexity, it requires more resources working on components integration. More resources means more money but most of Zentyal users are not Zentyal customers. They don't want to spend money buying support (which is current Zentyal business model) 

Although some Zentyal users are happy 3.2 users not facing that many bugs, this is perhaps not stable enough and should still be "beta". Do you really think this would prevent users to migrate, run this "beta" platform in prod and rant because this is not stable ? I don't think so 

[Lonniebiz]

@Christian: What file server do you recommend to use in a Zentyal domain environment? What are you using?

[soorploom]

Christian,

It matters not what is 'beta' inside z3.2 and whilst you are correct in that samba has changed significantly and moreover, the Zentyal implemention thereof, saying that Zentyal "has to..." is far, far too late. Now, if it was 'Zentyal had to...', then that would be a different matter entirely.

I'm not arguing the toss, at present, about Windows domain emulation or anything else although some of the other features appear to be on the flaky side too, simply the serving of files which is pretty much the raison d'etre of a server. I'm sure there is a large percentage of users, of whatever version, are utilising Zentyal as something akin to a NAS device. If that is the case, with the latest version, tough.

z3.2, I will say again, is being touted as a Microsoft SBS replacement. Not a beta replacement but a production replacement. Users will neither require nor understand why it seems necessary to have a different array of boxes to do other things, as in your comment that you personally use a different file server. You may have elected to do that but I doubt many others would.

It is a fact that file serving in z3.2 does not work.

A prospective user in particular will (practically) always obtain the advertised, existing version. Many will neither care nor be interested in previous versions probably on the assumption that the latest iteration of a product has been enhanced over the last and has been debugged to allow it to be all singing and all dancing but moreover, production. Not beta or anything else.

Could you imagine a conversation between a prospective customer and Zentyal sales, which might go something like, 'Yes, this works, so does that, this is new but, ehm, file sharing is in beta, it's a bit dodgy but we've released ten patches over the past few weeks so we're working on it'. Click, customer gone.

Zentyal, again as you put it, can 'rush' as much as it/they want. Fine but don't rush to announce a product as functional when clearly, it is not.

Like much in IT, especially something a bit different to the usual fare, it's only as good as its last success. Failure, in particular paid for failure, tends to get a reputation which sticks.

My own gripe, if that's the correct word, is the continual flow of patches for a product which is literally only weeks old and still it's unstable. Now, if there had been some guidelines, some feedback, indications as when this would be put to rights, fine. Well, for me it would be fine. Possibly not for others and I would assume definitely not for whatever number it is, who pay for support. Does this mean that those who are paying have the same problems or is it that the solution is available but only to paying customers? Unless I've missed something, I'm deafened by the silence from Zentyal on this forum.

If you consider this to be a "rant", OK. Maybe it is but nevertheless, z3.2 is not production as advertised, it may be, in truth beta but whatever its label, it does not work at least for those, as you say, who are not paying Zentyal customers.

We'll see what transpires over the coming days, weeks....whatever.

[christian]

Soorploom,

You're right.
I'm not suggesting to implement multiple boxes, I explain why I don't use Zentyal file sharing. My comment is perhaps misleading. Sorry for that.
I also do share that new comer expects to install last release and asking to deploy the previous one is not acceptable. Here again, my comment was toward current Zentyal users having migrated from 2.x to 3.0 then 3.2 but being disappointed with what they get.
This said, if Zentyal pushes lot of patches, one can expect to have problem solved. As long as there is no regression, I don't think this should be seen as negative point. Whatever Zentyal plan is, new Samba patch might be triggered by Samba team introducing changes or fixes 

My next comment is for sure biased as I'm not 3.2 user but I'm pretty sure some 3.2 users are currently quite happy with what they deployed. Same for 3.2 Zentyal customers.

Could you imagine a conversation between a prospective customer and Zentyal sales, which might go something like, 'Yes, this works, so does that, this is new but, ehm, file sharing is in beta, it's a bit dodgy but we've released ten patches over the past few weeks so we're working on it'. Click, customer gone.

You're perfectly right  if discussion is lead this way but I don't think this is the way it currently works. File sharing is not beta but evolving quickly.
What do you suggest here ? Not to follow Samba patches and publish patches only once or twice a year ?

For what I understand, main difference between community users and customers is not in term of features but in term of support plus, announced during last summit, the way releases and versions will be issued. customers will get less but more stable updates, once a year while community users will benefit from more updates (every quarter) but of course this is at the cost of potential stability issues.

Regarding your point about Zentyal being silent here... hum, this is community forum    you may try Zentyal IRC where perhaps more Zentyal staff use to join.

In parallel to this interesting debate about Zentyal strategy and support, you should, if not already done (which I doubt) create posts dedicated to technical problems you may face, just in case some other community user has an answer or fix for you.

[soorploom]

Christian,

Thank you for the super quick response, much more rapid than the last. I’m surprised but grateful.

“I'm not suggesting to implement multiple boxes, I explain why I don't use Zentyal file sharing.”

Mmmm, no, you didn’t suggest but you did imply that having a totally seperate system (dedicated) to file sharing is a workable situation, which, of course, it is and always would be assuming unnecessary complication is what you want. Of course it is, and given rather more resources than I would assume Zentyal’s core target clientele to be, technically the only way to go.

May I, in turn, suggest that you, again miss the point, the whole point, the reason for the whole and entire existence of Zentyal 3.2, is that it, ostensibly, does the lot and easily without delving into the multifarious guts of Ubuntu/linux, where no man-in-the-street dare venture.

Dare I suggest even further, that possibly better, smoother, cheaper than a rather larger, bigger and who-gives-a-toss-about-small-business-users IT company who could, quite easily, lose a whack of business in the direction of an up and coming bunch which has the right solution at the right time and, of course, at the right price with a correct community attitude. It is Ubuntu/linux, after all or has the ethos escaped me?

“My next comment is for sure biased as I’m not a 3.2 user”. Given what you say, how could you possibly be biased, one way or the other, if, as you state, you are not a user of z3.2? Indeed, with that in mind, how can you make any comment on z3.2 in any functioning environment?

You further add, “File sharing is not beta but evolving quickly”. So, is that something like, ‘sitting on an aircraft, hundreds of souls ready for the off only to be told that it all is fine, except that the design of the wings is evolving quickly (but as of now, still not there)? .....hundreds of souls then run like hell off that plane.

Have to say that, “File sharing is not beta but evolving quickly”, in my simple parlance means, it’s fu…broken, doesn’t work but maybe it will, sometime in the future. Given that, all promises are always in the future and all lies and mistakes are in the past, that fills me with unbridled optimism, especially in an IT environment where optimism always comes to fruition and strangely, always but always in the future.

You further state, “For what I understand, main difference between community users and customers is not in term of features but in term of support…”.

I think you understand all too well. However, if it is, as you allude, unless I caught this obtusely, customers get it all as in a working model but community users pick up what they can get, fine. I won’t go any further except to say that Zentyal’s advertising stuff doesn’t bring that point into focus. At all.

Yes, this *is* a community forum. Note, ‘community’. I thought that’s where all issues were brought to the fore so that the ‘community’ could deal with issues, etc. Even bad ones.

You might also be aware that myself and others have posted z3.2 issues dedicated to technical problems we *do* face. Others have attempted to address the failure of z3.2, in the particular case of samba but as I can see thus far, with no success, even after ten patches.

[mpnegro]

Hi,
I'm a "community" and a "costumer", The main difference is support, only that. I replicated the office server in my home server, to test implementations and stuffs like that.
Soo when 3.2 gets out, I wait 1 week and update my home server. Obviously, I'm like most of the 3.2 users, with a semi-functional file server. At the office we don't update because of this and after some talk with Zentyal staff.
I agree that 3.2 is not ready for "prime time" but they are working on it.
Your point is correct, they should not have make this release as ready but as a beta.
Some of the Zentyal people, like Javi are answer some threads (i see some answers from him in the English forum, Spanish and Portuguese the other languages i don't know), so Zentyal staff (some of them) talk to the users. In ubuntu forums I never catch a dev answer.

Make the thing short, 3.2 is beta at  the most, and they "blunder" big time. Now what we can do? Like always try to fill the best bugs we can and try to find workarounds.
Sorry this are my 2 cents.
Miguel

[christian]

“My next comment is for sure biased as I’m not a 3.2 user”. Given what you say, how could you possibly be biased, one way or the other, if, as you state, you are not a user of z3.2? Indeed, with that in mind, how can you make any comment on z3.2 in any functioning environment?

You further add, “File sharing is not beta but evolving quickly”. So, is that something like, ‘sitting on an aircraft, hundreds of souls ready for the off only to be told that it all is fine, except that the design of the wings is evolving quickly (but as of now, still not there)? .....hundreds of souls then run like hell off that plane.

   We can discuss forever, this will for sure lead nowhere as we are definitely not on the same track. You expect open source based on the very last Samba 4 version to be rock stable while I don't understand how this could be achieve. Period.

Then one could discuss if 3.2 should be beta or not. I don't want to take part to this useless (from my standpoint) debate.

If your requirement is rock stable platform, the right choice is definitely not the very last version of any software based on quickly evolving component. If this is not clear to you, this is at least to me and not related to Zentyal only. In almost all IT services, worldwide, when new software release is published, IT staff spends some time ensuring that it works as expected and that migration will not impact service continuity.
It looks like behaviour is different with Zentyal community given the amount of messages here stating: "please help, I've migrated my production platform to 3.2 and it doesn't work any more"

Here again, my point is not to discuss if 3.2 is supposed to work because this is no more beta. Yes it should work, I do, fully, agree. Obviously, it doesn't, at least for all users. What do you suggest to help with such situation ?

You further state, “For what I understand, main difference between community users and customers is not in term of features but in term of support…”.

I think you understand all too well. However, if it is, as you allude, unless I caught this obtusely, customers get it all as in a working model but community users pick up what they can get, fine.

Not being Zentyal customer, I don't want to make wrong statement here regarding evolution that has been announced during last summit. So far, difference is in term of support and I suppose, well I should rather said "I hope" that Zentyal official support is able to help customers achieving deployment of stable platform even on 3.2. In term of feature, I don't think there is any difference.
However I'm not surprised that forum not being able to provide same level of support than Zentyal support.

You might also be aware that myself and others have posted z3.2 issues dedicated to technical problems we *do* face. Others have attempted to address the failure of z3.2, in the particular case of samba but as I can see thus far, with no success, even after ten patches.

Reading this forum quite a lot  I do know that 3.2 brought and still brings a lot of problem and frustration. I do not try to find excuses for Zentyal and I don't want to argue forever whether 3.2 should be stamped as release or beta.
If you need stable platform, why don't you revert back to previous stable version while, in parallel, you could spend more than 2 posts among the 10 you already posted to investigate on the technical aspects of failure or problems you face ?

TTFN.

[soorploom]

Christian,

I thought I would sit on the side for a few days in the desperate hope that Zentyal would magically pop out another, yet another, patch which would allow z3.2 to actually share files.

Lo and behold, yet another upgrade to samba4. Getting a bit transfixed with the number and frequency of patches but I think that was available from maybe 9 December 2013? Excitement! At last. A server actually being capable of serving files might be within reach.

Err, no. Blinded by the 67% gremlin, yet again. Does not instil hope but then again, that was starved of confidence about five or six patches ago.

The simple fact is that Zentyal 3.2 is being touted as the be all and end all to those who, hitherto, may have elected to go the Microsoft SBS route. I assume that you have perused the zentyal.com site, not that it gives any in depth detail.

You have constantly attempted to defend Zentyal. You have constantly made excuses. You have suggested that it might be better to return to a previous version, which in itself, immediately poses questions as to the integrity of the currently offered version.

"If you need stable platform, why don't you revert back to previous stable version" That looks strangely like a comment from someone with a close connection to Zentyal corporate.

The simple fact is, following the release of z3.2 and the multitude of patches, Zentyal 3.2 does not correctly serve files, if at all. Hence, it does not work. It is and remains broken. A server, irrespective of its other claimed abilities, which is incapable of serving files, is no server.

As a consequence, at best, Zentyal 3.2 is extremely fragile. At worst, time consuming and totally useless.

Until Zentyal fixes this release, further conversation is pointless.

[christian]

You have constantly attempted to defend Zentyal. You have constantly made excuses. You have suggested that it might be better to return to a previous version, which in itself, immediately poses questions as to the integrity of the currently offered version.

"If you need stable platform, why don't you revert back to previous stable version" That looks strangely like a comment from someone with a close connection to Zentyal corporate.

Apologize. My fault    I initially though your ultimate goal was to implement stable file sharing service, reason why I offered alternative approaches.
If goal is to discuss about 3.2 only, I already explained that I'm not running it, for multiple reasons, one being Zentyal decision to rely on almost still beta Samba 4 
That said, no, I'm not close to Zentyal otherwise I would definitely promote 3.2  which I don't

The simple fact is, following the release of z3.2 and the multitude of patches, Zentyal 3.2 does not correctly serve files, if at all. Hence, it does not work. It is and remains broken. A server, irrespective of its other claimed abilities, which is incapable of serving files, is no server.

As a consequence, at best, Zentyal 3.2 is extremely fragile. At worst, time consuming and totally useless.

Until Zentyal fixes this release, further conversation is pointless.

I do agree