Well, a lot happened since Telligent aquired Zimbra.
BUT: the source tree still doesnt compile without dozens of patches.
And just have a look at their reactions on community input in bugzilla. For me it now seems pretty clear that they are NOT interested in any contributions at all - the whole OSS topic seems to be nothing but an alibi.
The End of Zimbra
The End of Zimbra
A few more points:
1. We're actually converting the Zimbra Community (formerly Telligent) to use MariaDB and have been a long time user of Apache Solr in the Zimbra Community product.
2. The technology used by Zimbra Community has no impact what-so-ever on the technology used by Zimbra Collaboration (or for that matter Zimbra Sync and Share which is built on Python). Any and all integration that we've done or talked about is done at the API level. There is no "mass conversion" to a single technology.
3. Yes, all of our products are designed to be white labeled and all of our products are designed for large scale implementations (all of the products are multi-tenant).
4. Roadmap compared to VMWare? So much to cover, but we are executing against all of the commitments (or nearly all) that were communicated on while Zimbra was ignored at VMWare. We also brought back Zimbra Desktop and are investing in that product again - something that had been killed.
Happy to answer any questions about our products or strategy anytime: rob@zimbra.com
1. We're actually converting the Zimbra Community (formerly Telligent) to use MariaDB and have been a long time user of Apache Solr in the Zimbra Community product.
2. The technology used by Zimbra Community has no impact what-so-ever on the technology used by Zimbra Collaboration (or for that matter Zimbra Sync and Share which is built on Python). Any and all integration that we've done or talked about is done at the API level. There is no "mass conversion" to a single technology.
3. Yes, all of our products are designed to be white labeled and all of our products are designed for large scale implementations (all of the products are multi-tenant).
4. Roadmap compared to VMWare? So much to cover, but we are executing against all of the commitments (or nearly all) that were communicated on while Zimbra was ignored at VMWare. We also brought back Zimbra Desktop and are investing in that product again - something that had been killed.
Happy to answer any questions about our products or strategy anytime: rob@zimbra.com
The End of Zimbra
Looking forward to the MariaDB backend for Zimbra Community. Great information [mention:ca8fea2575d649a9926b149790a90b7d:e9ed411860ed4f2ba0265705b8793d05]
The End of Zimbra
[QUOTE]
Zimbra Collaboration Server is still the same great email and collaboration product it's always been with the very same,
[/QUOTE]
Yeah, and the codebase is as broken as it has ever been. That thing does not even compile.
And there're not even any SCm tags to the source code of specific releases.
I've filed lots of bug reports and submitted fixes, all of them got rejected with silly reasons (or no comment at all).
Looking at what's happening in bugzilla, it _really_ seems that you folks simply do not want anybody out here
being able to compile that thing on our own, assuming everybody should just use the precompiled binaries,
an ABSOLUTE NO-GO for my projects. That _breaks_ the whole idea of OpenSource.
It really looks like the whole OSS topic here is nothing but an marketing bluff.
[QUOTE]
very talented developers
[/QUOTE]
Well, so talented on regularily breaking the codebase, introducing new regressions every day ?
Too talented for being able (or willing) to even understand the actual problems mentioned in bug reports ?
Too talented for using the SCM properly ? (or even *having* a proper SCM ... and no: P4 is totally crap!)
Too talented for beng able to write some _actually working_ build scripts ?
I'm closely following the changes in the codebase for several years now. And sorry folks, it's really
of terrible quality. For a short time, it seemed to become better, but in recent weeks it's again
getting worse day per day. The latest move to maven (made by dumb people for even more dumb people)
introduced so horrible breaks, that it's now completely unusable.
Congratulations! You guys really managed to completely mess up a formerly good product.
[QUOTE]
none of that will be changing
[/QUOTE]
If none of that will be changing, I see no future at all for Zimbra.
Meanwhile, I have to propose all my clients to drop it and move to Kolab.
(by the way: you know why certain large German state capital picked Kolab instead of Zimbra ?)
[QUOTE]
(except lots of new exciting features)
[/QUOTE]
"exciting" features are worth nothing, if the whole thing is completely broken and operating
is expensive because you didn't even get it properly packaged.
Zimbra Collaboration Server is still the same great email and collaboration product it's always been with the very same,
[/QUOTE]
Yeah, and the codebase is as broken as it has ever been. That thing does not even compile.
And there're not even any SCm tags to the source code of specific releases.
I've filed lots of bug reports and submitted fixes, all of them got rejected with silly reasons (or no comment at all).
Looking at what's happening in bugzilla, it _really_ seems that you folks simply do not want anybody out here
being able to compile that thing on our own, assuming everybody should just use the precompiled binaries,
an ABSOLUTE NO-GO for my projects. That _breaks_ the whole idea of OpenSource.
It really looks like the whole OSS topic here is nothing but an marketing bluff.
[QUOTE]
very talented developers
[/QUOTE]
Well, so talented on regularily breaking the codebase, introducing new regressions every day ?
Too talented for being able (or willing) to even understand the actual problems mentioned in bug reports ?
Too talented for using the SCM properly ? (or even *having* a proper SCM ... and no: P4 is totally crap!)
Too talented for beng able to write some _actually working_ build scripts ?
I'm closely following the changes in the codebase for several years now. And sorry folks, it's really
of terrible quality. For a short time, it seemed to become better, but in recent weeks it's again
getting worse day per day. The latest move to maven (made by dumb people for even more dumb people)
introduced so horrible breaks, that it's now completely unusable.
Congratulations! You guys really managed to completely mess up a formerly good product.
[QUOTE]
none of that will be changing
[/QUOTE]
If none of that will be changing, I see no future at all for Zimbra.
Meanwhile, I have to propose all my clients to drop it and move to Kolab.
(by the way: you know why certain large German state capital picked Kolab instead of Zimbra ?)
[QUOTE]
(except lots of new exciting features)
[/QUOTE]
"exciting" features are worth nothing, if the whole thing is completely broken and operating
is expensive because you didn't even get it properly packaged.
The End of Zimbra
[QUOTE]
Zimbra have made a commitment to remain open source for the ZCS products.
[/QUOTE]
Bill, do you know what's the core idea of OpenSource ?
It's primarily about:
a) everybody being able to just fetch the source and build it on his own
b) open collaboration between people all around the world
Exactly the two points that Zimbra FAILS to meet - for many years now. And it doesn't seem to get better - it's getting worse.
Just look at what's happening in bugzilla. Your guys there clearly DO NOT WANT any community input !
The whole "OSS" blablah here is nothing but marketing junk. Cannot be taken seriously.
Zimbra have made a commitment to remain open source for the ZCS products.
[/QUOTE]
Bill, do you know what's the core idea of OpenSource ?
It's primarily about:
a) everybody being able to just fetch the source and build it on his own
b) open collaboration between people all around the world
Exactly the two points that Zimbra FAILS to meet - for many years now. And it doesn't seem to get better - it's getting worse.
Just look at what's happening in bugzilla. Your guys there clearly DO NOT WANT any community input !
The whole "OSS" blablah here is nothing but marketing junk. Cannot be taken seriously.
The End of Zimbra
I understand you want things to move faster.
Some of this is unwinding what happened at VMWare. For example, while it seems simple, we made a pretty significant change a few months back moving from the proprietary "ZPL" (Zimbra Public License) to the GPL v2. And we have been working through, for several months now, on how to enable the community to contribute to the code base.
We are also balancing that against the commercial interests of the business, which enable us to keep the lights on, etc.
Some of this is unwinding what happened at VMWare. For example, while it seems simple, we made a pretty significant change a few months back moving from the proprietary "ZPL" (Zimbra Public License) to the GPL v2. And we have been working through, for several months now, on how to enable the community to contribute to the code base.
We are also balancing that against the commercial interests of the business, which enable us to keep the lights on, etc.
-
L. Mark Stone
- Ambassador
- Posts: 2802
- Joined: Wed Oct 09, 2013 11:35 am
- Location: Portland, Maine, US
- ZCS/ZD Version: 10.0.7 Network Edition
- Contact:
The End of Zimbra
Rob,
When we met at last year's Big Social you made a big deal about being strict with communication happening at the API level. Having headed a $5 million software development project years ago your statement got big points from me. And I agree that in one short year Telligent (nee Zimbra) have made great strides with freeing Zimbra from VMware.
Hindsight is always 20/20 but if it had been my call I would have brought back IM before resurrecting Zimbra Desktop. Especially because certain Zimbra employees at last year's Big Social indicated IM would be back in Zimbra 8.5, so we updated a number of our clients to 8.0.x in anticipation of 8.5 including IM only to be disappointed. Had we known, we would have kept folks on 7.2.x.
pm.zimbra.com continues to be broken; 8.0.8 went GA in the past few days but pm.zimbra.com shows 8.0.8 nowhere near ready to be released.
I think the "Big Picture" strategy emphasis on things like "Always On" architecture is right on; but my sense is a number of the smaller bt quite important "blocking and tackling" items are not getting their due. Our hosting environment has fully redundant networking, N+x compute nodes, and redundant SAN storage. We already domicile tertiary Zimbra MX servers in different data centers and have tested Zimbra mailbox servers in data centers different from where the proxy server(s) are located. Our clients are pretty annoyed at the loss of IM and not seeing/appreciating the benefits of "Always On" architecture when our hosting environment underneath Zimbra is already quite redundant/resilient.
Release Notes continue to have inaccuracies in them, causing admins to scramble to do things like install prerequisite operating system packages not listed in the documentation -- only after the Zimbra installer exists after reporting missing dependencies documented nowhere except in bugzilla (and Zimbra is already down...).
An in-place upgrade of Ubuntu from 10.04 to 12.04 or 14.04 will similarly fail if users run the installer-prompted MySQL database check (same version of Zimbra). Not documented anywhere except, again, in bugzilla.
Again, I like the strategic direction, but I feel that a lot of execution elements are getting less attention and causing us that host and support Zimbra for customers to spend more time apologizing for Zimbra than we have had to and would like to.
Don't take your foot off the throttle, but don't go too long between pit stops either -- you know what happens when you let your tires get too worn!
All the best,
Mark
When we met at last year's Big Social you made a big deal about being strict with communication happening at the API level. Having headed a $5 million software development project years ago your statement got big points from me. And I agree that in one short year Telligent (nee Zimbra) have made great strides with freeing Zimbra from VMware.
Hindsight is always 20/20 but if it had been my call I would have brought back IM before resurrecting Zimbra Desktop. Especially because certain Zimbra employees at last year's Big Social indicated IM would be back in Zimbra 8.5, so we updated a number of our clients to 8.0.x in anticipation of 8.5 including IM only to be disappointed. Had we known, we would have kept folks on 7.2.x.
pm.zimbra.com continues to be broken; 8.0.8 went GA in the past few days but pm.zimbra.com shows 8.0.8 nowhere near ready to be released.
I think the "Big Picture" strategy emphasis on things like "Always On" architecture is right on; but my sense is a number of the smaller bt quite important "blocking and tackling" items are not getting their due. Our hosting environment has fully redundant networking, N+x compute nodes, and redundant SAN storage. We already domicile tertiary Zimbra MX servers in different data centers and have tested Zimbra mailbox servers in data centers different from where the proxy server(s) are located. Our clients are pretty annoyed at the loss of IM and not seeing/appreciating the benefits of "Always On" architecture when our hosting environment underneath Zimbra is already quite redundant/resilient.
Release Notes continue to have inaccuracies in them, causing admins to scramble to do things like install prerequisite operating system packages not listed in the documentation -- only after the Zimbra installer exists after reporting missing dependencies documented nowhere except in bugzilla (and Zimbra is already down...).
An in-place upgrade of Ubuntu from 10.04 to 12.04 or 14.04 will similarly fail if users run the installer-prompted MySQL database check (same version of Zimbra). Not documented anywhere except, again, in bugzilla.
Again, I like the strategic direction, but I feel that a lot of execution elements are getting less attention and causing us that host and support Zimbra for customers to spend more time apologizing for Zimbra than we have had to and would like to.
Don't take your foot off the throttle, but don't go too long between pit stops either -- you know what happens when you let your tires get too worn!
All the best,
Mark
___________________________________
L. Mark Stone
Mission Critical Email - Zimbra VAR/BSP/Training Partner https://www.missioncriticalemail.com/
AWS Certified Solutions Architect-Associate
L. Mark Stone
Mission Critical Email - Zimbra VAR/BSP/Training Partner https://www.missioncriticalemail.com/
AWS Certified Solutions Architect-Associate
The End of Zimbra
[quote]
I understand you want things to move faster.
[/quote]
No. I want things to be done CAREFUL. First stabilize the current codebase (_including_ the build, of course!)
and fix bugs, before adding any fancy new features. (you probably know, I've already done lots of Zimbra
projects with deep customizations - stability and reproducability is the PRIMARY requirement for serious
enterprise usecases - all the fancy features are just useless, if the application doesn't work well).
In fact, both things could be easily done in parallel - if you had a decent SCM (and people who know how to use it),
with that silly Perforce monster, you're completely out of luck - it's just unusable crap.
For all my clients, the first things I usually do is: introduce a proper SCM infrastructure (which also includes
proper workflows, of course), CI and packaging. As long as the build doesn't work automatically and reliably,
any other developments are quite useless.
[quote]
For example, while it seems simple, we made a pretty significant change a few months back moving from the
proprietary "ZPL" (Zimbra Public License) to the GPL v2.
[/quote]
That's fine, but wasn't that important, as long as the whole codebase doesn't even build.
[quote]
And we have been working through, for several months now, on how to enable the community to contribute
to the code base.
[/quote]
Fine, and what was the practical outcome ?
Okay, after years of empty promises, we now at least have some (yet incomplete) mirror to git. But still
the tags - which are _vital_ - are still missing. So we (the community - if there is any) have no way to
get the exact revisions for certain releases.
By the way: I already tried to contribute lots of things. But everything (in fact, pretty much my bug reports)
got refused with silly excuses. In general, your guys just deny that there are problems at all. Just have a
look at bugzilla ...
[quote]
We are also balancing that against the commercial interests of the business, which enable us to keep the lights on, etc.
[/quote]
Well, not making fool of yourself, by exposing yourself as completely unorganized and frouzy, ignoring
customer and community needs, sticking with silly and broken concepts, etc, etc, *is* an business
interest - isn't it ?
I'd guess, working cost-efficient should be one, too. Well, with your current methodologies (not even
having a decent SCM-,CI-,build-/packaging infrastructure), you just waste a lot of resources.
Yes: I've been through this several times - by implementing such things, I usually reduce the development
costs for my clients in magnitudes. (and yes: that's also why I developed ZMPKG - something I already
offered _years_ ago).
I understand you want things to move faster.
[/quote]
No. I want things to be done CAREFUL. First stabilize the current codebase (_including_ the build, of course!)
and fix bugs, before adding any fancy new features. (you probably know, I've already done lots of Zimbra
projects with deep customizations - stability and reproducability is the PRIMARY requirement for serious
enterprise usecases - all the fancy features are just useless, if the application doesn't work well).
In fact, both things could be easily done in parallel - if you had a decent SCM (and people who know how to use it),
with that silly Perforce monster, you're completely out of luck - it's just unusable crap.
For all my clients, the first things I usually do is: introduce a proper SCM infrastructure (which also includes
proper workflows, of course), CI and packaging. As long as the build doesn't work automatically and reliably,
any other developments are quite useless.
[quote]
For example, while it seems simple, we made a pretty significant change a few months back moving from the
proprietary "ZPL" (Zimbra Public License) to the GPL v2.
[/quote]
That's fine, but wasn't that important, as long as the whole codebase doesn't even build.
[quote]
And we have been working through, for several months now, on how to enable the community to contribute
to the code base.
[/quote]
Fine, and what was the practical outcome ?
Okay, after years of empty promises, we now at least have some (yet incomplete) mirror to git. But still
the tags - which are _vital_ - are still missing. So we (the community - if there is any) have no way to
get the exact revisions for certain releases.
By the way: I already tried to contribute lots of things. But everything (in fact, pretty much my bug reports)
got refused with silly excuses. In general, your guys just deny that there are problems at all. Just have a
look at bugzilla ...
[quote]
We are also balancing that against the commercial interests of the business, which enable us to keep the lights on, etc.
[/quote]
Well, not making fool of yourself, by exposing yourself as completely unorganized and frouzy, ignoring
customer and community needs, sticking with silly and broken concepts, etc, etc, *is* an business
interest - isn't it ?
I'd guess, working cost-efficient should be one, too. Well, with your current methodologies (not even
having a decent SCM-,CI-,build-/packaging infrastructure), you just waste a lot of resources.
Yes: I've been through this several times - by implementing such things, I usually reduce the development
costs for my clients in magnitudes. (and yes: that's also why I developed ZMPKG - something I already
offered _years_ ago).
The End of Zimbra
[quote]
Hindsight is always 20/20 but if it had been my call I would have brought back IM before resurrecting Zimbra Desktop.
[/quote]
Actually, I don't think that Zimbra should bring it's own IM server, but instead integrate well with an existing
one (eg. openfire), including an chat xmpp-based zimlet.
OTOH, such features are already are provided by other vendors, such as VNC.
[quote]
Especially because certain Zimbra employees at last year's Big Social indicated IM would be back in Zimbra 8.5, so we updated a number of our clients to 8.0.x in anticipation of 8.5 including IM only to be disappointed. Had we known, we would have kept folks on 7.2.x.
[/quote]
Quite understandable. Perhaps give VNC Chat Zimlet a try.
[quote]
pm.zimbra.com continues to be broken; 8.0.8 went GA in the past few days but pm.zimbra.com shows 8.0.8 nowhere near ready to be released.
[/quote]
Actually, I can't remember a day when it wasn't broken or outdated.
@Rob: why don't you just generate it from Bugzilla ? Shouldn't take much more than a few lines of PHP code.
Or, if your community product is so fine as you're always promote, it shouldn't be a big deal to integrate it there.
Of course, that would require your devs to actually maintain bugzilla. I see many thousands open bugs (!!!) there.
Seriously, are all these issues really still valid ?
[quote]
I think the "Big Picture" strategy emphasis on things like "Always On" architecture is right on; but my sense is a number of the smaller bt quite important "blocking and tackling" items are not getting their due.
[/quote]
Seconded. I'd really advise first stabilizing the existing codebase before introducing fancy new things.
[quote]
Release Notes continue to have inaccuracies in them, causing admins to scramble
[/quote]
@Rob: if you would have a decent SCM (and people actually using it in a sane way), you could auto-generate
a usable change list, upon which the documentation team could easily update/review the documents without
easily missing anything. And, of course, sane issue tracking workflows/policies would be very helpful here.
Yet again, clearly a lack of proper development management. You need an seasoned senior devops manager.
[quote]
to do things like install prerequisite operating system packages not listed in the documentation -- only after the Zimbra installer exists after reporting missing dependencies documented nowhere except in bugzilla (and Zimbra is already down...).
[/quote]
This installer is conceptionally wrong in the first place. Instead, the correct way would be properly
packaging the application for the corresponding platforms, so their package management infrastructure
handles everything. The operator should only have to add the Zimbra repo to his system and fire up
an 'apt-get install ...' / 'yum install ...' command.
I've raised that issue _years_ ago. Of course, it was rejected, with silly excuses.
[quote]
An in-place upgrade of Ubuntu from 10.04 to 12.04 or 14.04 will similarly fail if users run the installer-prompted
MySQL database check (same version of Zimbra). Not documented anywhere except, again, in bugzilla.
[/quote]
Well, having that reported @bugzilla is correct, as this *IS* a bug. Indeed, a really critical one.
Of course, known issues and proper workarounds have to be documented properly !
BUT: WTF isn't that bug still not fixed, while resources are spent for the next fancy new (mis)feature
or making the build system even more broken by introducing ridiculus things like maven ?!
[quote]
Again, I like the strategic direction, but I feel that a lot of execution elements are getting less attention and
causing us that host and support Zimbra for customers to spend more time apologizing for Zimbra than
we have had to and would like to.
[/quote]
I know how you're feeling. Some time ago, we regularily had to do lots of work arounds for pretty
critical security bugs, while Zimbra folks just ignored the issues. And these were (full) NE customers
with many thousands of mailboxes (I'd guess they payed several 100k$ for licenses only).
Hindsight is always 20/20 but if it had been my call I would have brought back IM before resurrecting Zimbra Desktop.
[/quote]
Actually, I don't think that Zimbra should bring it's own IM server, but instead integrate well with an existing
one (eg. openfire), including an chat xmpp-based zimlet.
OTOH, such features are already are provided by other vendors, such as VNC.
[quote]
Especially because certain Zimbra employees at last year's Big Social indicated IM would be back in Zimbra 8.5, so we updated a number of our clients to 8.0.x in anticipation of 8.5 including IM only to be disappointed. Had we known, we would have kept folks on 7.2.x.
[/quote]
Quite understandable. Perhaps give VNC Chat Zimlet a try.
[quote]
pm.zimbra.com continues to be broken; 8.0.8 went GA in the past few days but pm.zimbra.com shows 8.0.8 nowhere near ready to be released.
[/quote]
Actually, I can't remember a day when it wasn't broken or outdated.
@Rob: why don't you just generate it from Bugzilla ? Shouldn't take much more than a few lines of PHP code.
Or, if your community product is so fine as you're always promote, it shouldn't be a big deal to integrate it there.
Of course, that would require your devs to actually maintain bugzilla. I see many thousands open bugs (!!!) there.
Seriously, are all these issues really still valid ?
[quote]
I think the "Big Picture" strategy emphasis on things like "Always On" architecture is right on; but my sense is a number of the smaller bt quite important "blocking and tackling" items are not getting their due.
[/quote]
Seconded. I'd really advise first stabilizing the existing codebase before introducing fancy new things.
[quote]
Release Notes continue to have inaccuracies in them, causing admins to scramble
[/quote]
@Rob: if you would have a decent SCM (and people actually using it in a sane way), you could auto-generate
a usable change list, upon which the documentation team could easily update/review the documents without
easily missing anything. And, of course, sane issue tracking workflows/policies would be very helpful here.
Yet again, clearly a lack of proper development management. You need an seasoned senior devops manager.
[quote]
to do things like install prerequisite operating system packages not listed in the documentation -- only after the Zimbra installer exists after reporting missing dependencies documented nowhere except in bugzilla (and Zimbra is already down...).
[/quote]
This installer is conceptionally wrong in the first place. Instead, the correct way would be properly
packaging the application for the corresponding platforms, so their package management infrastructure
handles everything. The operator should only have to add the Zimbra repo to his system and fire up
an 'apt-get install ...' / 'yum install ...' command.
I've raised that issue _years_ ago. Of course, it was rejected, with silly excuses.
[quote]
An in-place upgrade of Ubuntu from 10.04 to 12.04 or 14.04 will similarly fail if users run the installer-prompted
MySQL database check (same version of Zimbra). Not documented anywhere except, again, in bugzilla.
[/quote]
Well, having that reported @bugzilla is correct, as this *IS* a bug. Indeed, a really critical one.
Of course, known issues and proper workarounds have to be documented properly !
BUT: WTF isn't that bug still not fixed, while resources are spent for the next fancy new (mis)feature
or making the build system even more broken by introducing ridiculus things like maven ?!
[quote]
Again, I like the strategic direction, but I feel that a lot of execution elements are getting less attention and
causing us that host and support Zimbra for customers to spend more time apologizing for Zimbra than
we have had to and would like to.
[/quote]
I know how you're feeling. Some time ago, we regularily had to do lots of work arounds for pretty
critical security bugs, while Zimbra folks just ignored the issues. And these were (full) NE customers
with many thousands of mailboxes (I'd guess they payed several 100k$ for licenses only).