[SOLVED] I want to downgrade - 3.0.x community edition is not suitable ...

[thorsten]

OK,

I just started 10 days ago to upgrade to 3.0.x and I encountered too much bugs and not working things:

My personal dictionary reads for "UPGRADE":

Change old well known bugs against fresh unknown bugs ...  This is especially valid for Zentyal

Does anybody have a link to the latest 2.2.x version? This was much more stable and reliable for productive environment.

Just one question: I thougt - but I am not sure, the only difference between community and commercial edition is just the support and updates are tested. This means that 3.0 is equal for both edition at release, correct? But is the commercial edition of Zentyal that buggy, too? From the experiences of all versions (1.4, 2.x and 3.) I had before I will defiantly not use Zentyal for business purposes, this is too risky for me. As a playground for my home server it is quite nice, but as I do have a home office, I can not afford playing around with infrastructure issues. It is much safer to run a dedicated ADSL modem / router plus NAS than Zentyal.

Sorry, I am really frustrated at the moment - this upgrade is a deja vu for 2.0 .... It took me until december until I found out how to set up a bridged network which should work out of the box. Again it is december and I try to make thinks run that should work out of the box ...

So again: Is there some download link to 2.2.x

TIA
Thorsten

[Sam Graf]

I think 3.0 has been "betrayed" by a variety of things and remain optimistic long term. But I understand the frustrations.

Sourceforge has the old versions:

http://sourceforge.net/projects/zentyal/files/Old%20versions/

[Marcus]

Hello thorsten,

I'm maintaining the old, still stable, v2.2 installation documentation. You'll find 2.2 download links in it...
http://trac.zentyal.org/wiki/Documentation/Community/French/Installation/InstallationGuide

Direct links:
Zentyal 2.2-2 32bit installer 
https://sourceforge.net/projects/zentyal/files/zentyal-2.2-2-i386.iso/download

md5
https://sourceforge.net/projects/zentyal/files/zentyal-2.2-2-i386.iso.md5/download


Zentyal 2.2-2 64bit installer 
https://sourceforge.net/projects/zentyal/files/zentyal-2.2-2-amd64.iso/download

md5
https://sourceforge.net/projects/zentyal/files/zentyal-2.2-2-amd64.iso.md5/download


Best,

Marcus

[thorsten]

Dear Sam,

lets run into some philosophic discussion: 

I think 3.0 has been "betrayed" by a variety of things and remain optimistic long term.

I think you are right - but this is a standard problem for any product - marketing people create a promise for their product which is not kept. The strange thing is, that the promoise is not given oraly or by written letter, it is simply an "image" in the customers mind. E.g. smoking severe cigarettes shall be adventurous, using a special parfume is sexy and driving a special car is sportive. Such promises may concern product quality in termes of "suitability for intended purpose", quality in termes of "long lasting reliability & free from errors", features, design, (user friendly) usability etc.

If the product does not meet a certain quotient of inherently promises given, the custumer is disappointed. In the case of Zentyal, the features and comprehensive functionality is unbeatable e.g. compered to CLEAR OS. In contrast to this, version 3.0 is obviously buggy. Now the problem of the customer is, that he will extrapolate a certain product experiance to a brand image: If you are not satisfied with the quality (in terms of "free from errors") of a VW Beetle you will not buy a VW Phaeton, even without knowing that specific car. 

So back to the point: I think one inherent promise, formulated by the software market itself (NOT formulated by any company) is: A higher version number means more stability -  more errors are removed as the programmes have learned. Rembember Windows 3.11, Win 95, Win98 - think of all the blue sceens. I do not even remember tho have any blue sceen on my very stable XP and Win7 seems even better. I am sorry, but I expect this from increasing version numbers of Zentyal, too.

Window7 made as mistakte compared to Win XP: It is more stable, but functions need to be turned on / of at a different position. This is not good or bad, but is request the user to change mind on where to find options. New users may even find theirself much faster and from a objective point of view, intuitive usability may be even increased. But this is different for old users. Its like entrering a supermarkt 100 times a year - after the sales clerk did any rearrangement, you need to search for the cans you are used to. Additionally, for LINUX servers, you do not always know if you are right, sometimes you simply do a trail and error approach.

Link those two thinks together and you will have a frustrated customer 

By the way - I think it is NOT the fault of the production / development people, I think the error is done by the markting. 
 Typically a DILLBERT-like marketing pharse is:
"There is a complain, your product does not meet the customers requirements & expectations. Production shall perform better, we will need daily status reports on way you are so behind" 
But lets put it vice versa:
"The complain is, because the customer has the wrong promise in his mind. The wrong product features / usability / stability / timelines / goals were promised for the produkt / project" 
But I also think this is not a marketing problem - you can not take all eventuallities / problems into account when planning such a complex project. I know, the customer wants a dead line for produtc availablity, but this is at least in my case not true:

I read the announcement and I was happy to have version 3.0 in September, so I started my calculations based on my experiences on 2.x last year:
Lets wait for December until must bugs are reported and solved. For me, I did not require version 3.0 in September, I would be happy to have it, but there is not need. 2.x was running (with small bugs) stable perfectly for me. So I would prefer to have the cool features of 3.0 released in Jan 2013 without bugs... (OK, in this case I would have waited until March for download). 

Best regards
Thorsten

[christian]

Interesting discussion you're launching right now.

I agree that 3.0 is not stable, at least not enough to go live.
I also share that this is not (only) because of technical mistakes. But this is not that simple. Even if marketing has some responsibility here, there are some other aspects to be taken in account, the main onces being, from my standpoint:
- commitment to provide 3.0 in time (compared to roadmap published long time ago in an hard-coded, almost way, at least in term of planning if not content)
- competition against other similar products (yes there is some business behind)
- (potential) customers always asking and pushing for more features.

When Zentyal announced what they planed to put within 3.0 release, I express how reluctant I was to rush so fast toward so many heavy changes. Still I understand why they do this. From marketing standpoint, this is almost mandatory. However, technical staff should have enough weight and influence in such company to warn marketing and not kill their own business with such big risky step forward.

I do not blame anyone here as I don't know where the right balance is between few small technical steps forwards, stable and under control but unhappy "customers" and fast rush ahead with attractive communication (and users still and always asking for more).

I'm a technician, so I'm pushing for reliability but what if you have stable reliable but non attractive product. You may win the battle but on the very long term only, when customers from competitors will come back to you while disappointed with unstable products. Where is your business here 

My last comment will be about your wise decision to wait a bit before moving to 3.0
Jump from 2.X to 3.0 is much bigger than 1.x to 2.0 and 2.0 to 2.x
It introduces many new features and some of them are relying on products that are still not "released" (like Samba 4) thus waiting only couple of month is obviously not enough 
Even when Samba 4 will be released I expect some heavy changes in Zentyazl design. Current one with multiple LDAP, multiple DNS and Kerberos servers is more, in my opinion, a draft to show features than final "design".
What I really feel funny here is that such advanced features provided by 3.0(like GPO support thanks to Samba 4) are not required by most of SMBs. They may want it because they always read a lot of publication on internet about domain and beauty of GPOs but is there any real benefit for them ?

Perhaps Zentyal is not hunting SMBs only  OK fine but in such a case, is current technical design suitable for medium to large companies ? 

[Sam Graf]

Since we are all just thinking out loud here, while I think about this more let me just jot down a couple of things.

First, from my point of view, Zentyal exists in a very different world from products like Windows (and its aging parent, OS/2). Consider Debian, where Zentyal really begins. Very few people that I know think of the upcoming Debian 7 primarily as a more stable version of Debian 6. I think the same is true of Ubuntu, Zentyal's immediate technological context. Leaving Ubuntu Desktop out of the conversation for the moment, people seem more often to think of Ubuntu 12.04 as an updated version of Ubuntu 10.04, not primarily as more stable.

I think the same is true of the underlying technologies. It seems the real goal is to add features while avoiding regressions. In some cases, the core technology is mature, and in others, the core technology is in transition, but in both cases, most new versions represent new features rather than increased stability. Hence Debian's highly cautious approach to upstream releases, for example.

So to me, Zentyla 3 starts with the understood need to follow the Ubuntu Server LTS cycle, and for similar reasons. Even if no new features are added, Zentyal  either has to parallel upstream activity or fork Ubuntu (preferably not a combination of both; binary compatibility with Ubuntu is in Zentyal's interests, it seems to me), or get left behind just in terms of the underlying technologies.

If people assume Zentyal 3 is primarily a more stable version of Zentyal 2 (and I don't deny that to be the assumption, and understandably, at least sometimes), then we have a "marketing" problem of a different kind, I think. For to me, that's an assumption more or less foreign to the context in which Zentyal is developed.

I'm going to stop there for now, but I need to make a confession of sorts. My view of this could be colored by experiences outside Zentyal. At home, I've tended to become more conservative when it comes to open source choices. My feeling is that a rush to new features in the open source world is complicating quality control at the distribution level. To me, newer versions are more likely to contain serious regressions than ever before in my experience. Just as a generalization, I've retreat to the safer harbors of slow-moving distributions. So even outside Zentyal, but still in the Linux world, I've stopped thinking of newer versions as more stable. Instead, I've gone to using newer distributions of older software (e.g., Debian).

[christian]

Sam,

I share your comment about Ubuntu not targeting more reliability from one version to another. That's the reason why I tend to prefer Debian 
To me, Unity is just waste of resources and, I agree, I do not expect 12.04 to be more stable than 10.10 but, on the other hand, I expect bugs solved between 10.10 and 12.04 not to be re-introduced again when 12.10 is issued.
My expectation regarding operating system is to get something reliable on top of which I can install software of my choice. Distribution is somewhat different.
Before 11.10, as I was not happy with Evolution, I just removed it and installed Thunderbird. What I mean to say with this is that Ubuntu can be widely customized. This is not that easy with Zentyal: if you are not happy with Apache as web server for Zentyal console, you should better move to another (competitor) product. And Zentyal is not a distribution you can customize but rather a whole product by itself relying on Ubuntu.

Another comment I would like to add is that market addressed by Zentyal does require reliability. This is mandatory. Assuming your comment about Ubuntu is the key point to accept potentially unstable platform, then they should quickly move to another OS. Who can seriously think about deploying internet gateway platform or file server for your company that would crash from time to time because someone developed crappy piece of code?

The only aspect that could make sense to me is that Zentyal community version is less polished than commercial one. This could be Zentyal's choice to push updates to community as see what happens. I'm not saying this is the right approach, neither saying Zentyal applies this but at least I could understand this.
I do understand that some Zentyal users have more than frustration due to their move to 3.0 if they run it in prod.

[Sam Graf]

Assuming your comment about Ubuntu is the key point to accept potentially unstable platform, then they should quickly move to another OS. Who can seriously think about deploying internet gateway platform or file server for your company that would crash from time to time because someone developed crappy piece of code?

The broad question is, do we have a technology problem (to the extent that we have a problem), or do we have a marketing problem? So far I'm thinking that if people generally assume Zentyal 3 to be a more stable version of Zentyal 2, then yes, we have a marketing problem, but not the kind I understand thorsten to be suggesting that we have.

Further, Zentyal 3 is the first step in aligning Zentyal with Ubuntu's current LTS release. I think that makes it impractical to think of Zentyal 3 as a more stable version of Zentyal 2, so thinking of it that way is going to be disappointing almost certainly. So again, if we have a marketing problem, it's along the lines of aligning customer expectations with certain realities that may rule out the "newer is better" mindset from a stability point of view.

That's more or less all I'm saying. So far. But looking ahead, the next step for me is to think more about the technology problem.

[christian]

yes we can cut this into slices and distinguish between marketing vs. technical problem. This is, frankly, interesting but I don't think most of users here care about this.
If you tell them "yes, it doesn't work but you also misunderstood our communication. There is so far no commitment to have it stable enough to be used in prod despite the 3.0 label", I doubt these Zentyal users will like your distinguish between marketing and technology, even if you are right 

I'm not saying Zentyal is faulty. I would rather say that users are faulty when pushing that hard to have new features and expect all these huge heavy changes to work smoothly in a couple of month while relying on RC (Samba)
So, yes there is perhaps some truncated marketing communication, or misunderstood, or whatever... my point here is that user moving their production platform from Zentyal to another solution due to disappointment with 3.0 will not come back easily to Zentyal and will even less easily move from "community user" to "customer", if you see what I mean.

Thus my point: if community version is different from standard version and if community users have to expect, by design, more problems or unreliability due to the way changes are pushed, then, yes, there is a marketing issue because it has to be stated in a clearer way.

[Sam Graf]

yes we can cut this into slices and distinguish between marketing vs. technical problem. This is, frankly, interesting but I don't think most of users here care about this.

Since thorsten brought it up as a frustrated user, I didn't think a couple of paragraphs or so to think about it was all that bad of an idea.

In other words, expectations matter, even if people don't care that they matter. It matters very much to simple small business technology directors and coordinators if Zentyal lives up to the expectation that it will take routine hand management out of the Linux server equation, even if no one cares about marketing versus technology.

My expectation regarding operating system is to get something reliable on top of which I can install software of my choice. Distribution is somewhat different.

This I'm not following. I would say that Zentyal is a respin of Unbuntu Server, a distribution, which is built on Debian, also a distribution, because it unites Linux and the Hurd (sort of) with other software to create a complete distribution, not just an OS. But I'm not interested in debating the terms, only in undertanding what you're trying to say.

Since Zentyal is fundamentally a Ubuntu server respin (like Mint, Kubuntu, etc. are respins), how faithful a respin is it? Is Zentyal introducing a Samba RC in spite of or becuase of Ubuntu? To me, jumping to saying "that users are faulty" skips too much undiscussed stuff in the middle. I don't see the value in that.

For in that case, Zentyal users are putting Zentyal under enough pressure for unwise tehcnical choices to be made, to the end that Zentyal users are going to go somewhere else for their solution. End of conversation, because Zentyal can't win. It can't meet user expectations either way.

Back to expectations, realistic ones and unrealistic ones. And how respins should or can realistically respond to both.

[christian]

For in that case, Zentyal users are putting Zentyal under enough pressure for unwise tehcnical choices to be made, to the end that Zentyal users are going to go somewhere else for their solution. End of conversation, because Zentyal can't win. It can't meet user expectations either way.

No, perhaps you can't but at least what you provide works 
If you can't because you don't have technical skills that your competitors have, too bad. But if you can't because you want to stay on the safe path not implementing unstable solution, then users moving to some more risky solutions may come back to Zentyal when you will be ready later.

On the long run, I think it's better to not attract customers due to lack of technical solution rather than having unhappy customers leaving because what you provide doesn't work.

Not willing to discuss "wording", I can't see Zentyal as a distribution. To me Zentyal is not yet just another distro but piece of code (meaning development behind) to provide technical platform based on third-party software. Like webmin with some more constraints (but also supposed to bring more reliability thanks to administration easiness), at least the way I see it.

[Sam Graf]

No, perhaps you can't but at least what you provide works 

Of course. I didn't think that was even up for question.

[Sam Graf]

The next thing I'm thinking about is the mental adjustments small business IT has to make to navigate a mixed approach to updates actually impressed on it from the outside world. It's possible to have equipment that cannot be updated at all (e.g., PC-based medical equipment--spooky), that can be updated pretty freely (e.g., Windows desktops, where the major update cycle is often hardware-driven--stuff quits working--in a small buisness environment), and equipment that should be updated with due care (pretty much everything else, including stuff accepting only firmware updates). In this last case small business IT gets closest to thinking (hopefully) like its enterprise big brother--that test environments should be part of business life.

I remember the first time eBox tanked--over a Squid or DansGuardian update if I recall correctly. Nobody expected it to happen, but the results were fairly serious: impacted servers and their networks could no longer connect to the Internet (again, if I remember right). That's kind of a nasty failure to recover from for ordinary small buisness people.

So I'm still thinking that Zentyal 3.0 has been "betrayed" by a number of things. It's not necessarily a "perfect storm" of bad stuff, but it does seem to me to be layered.

[half_life]

<SNIP>
It introduces many new features and some of them are relying on products that are still not "released" (like Samba 4) thus waiting only couple of month is obviously not enough 
Even when Samba 4 will be released I expect some heavy changes in Zentyazl design. Current one with multiple LDAP, multiple DNS and Kerberos servers is more, in my opinion, a draft to show features than final "design".
What I really feel funny here is that such advanced features provided by 3.0(like GPO support thanks to Samba 4) are not required by most of SMBs. They may want it because they always read a lot of publication on internet about domain and beauty of GPOs but is there any real benefit for them ?

<SNIP>

Side note :  Samba4 is released as of today.  It is no longer a beta or a release candidate.

[christian]

Side note :  Samba4 is released as of today.  It is no longer a beta or a release candidate.

Cool  Let's see how it helps Zentyal 

Frankly, this is a good news and I hope it will solve most of the problem 3.0 users currently face. This said, we were more having a debate about strategy and roadmap principles than specific technical point about Samba4, even if I personally that relying on RC was not the best idea.
As I wrote above, what has been achieved with current Zentyal 3.0 design has the big advantage to give an early taste of what Samba 4 can provide. Now that Samba 4 is released, I would not be surprised if Zentyal team refine their design.
Or, to be more transparent, I would be surprised if they don't, perhaps not quickly but at least on the middle term.

Take some breath and look at what Zentyal needs today in order to offer Samba 4 features aside other modules:
- 2 LDAP servers (which means synchronization)
- 2 DNS servers
- 2 Kerberos servers

Is it reasonable 

The quick idea in order to solve it is to rely on Samba 4 but if you do so, it means that Zentyal is, like Microsoft Windows, tightly linked to its "AD" equivalent backbone. You will need this "domain" approach even for simple internet gateway without file sharing. This puzzles me a bit and this is my point with the rush to Samba 4.

I suggest you have a look at this, paying attention to the "Active Directory Compatible Server" section.

OK, I've expressed my doubts and concerns. Now let's give Zentyal team some time to comment