[SOLVED] The end of Zimbra? update to 10 is impossible and 8 and 9 goes eol shortly ....

[oetiker]

Hi

What do you think ... Zimbra 10 has not implemented many necessary features and at the same time
Zimbra 8.8.15 and Zimbra 9 goes EOL... what can we do? Update to Zimbra 10 is impossible and not update as well...
.... I can't understand Synacore how can that be.. a company does this ... ??

End of General Support
Kindly note that Zimbra 9.0.0 and 8.8.15 will be reaching their end-of-life (EOL). Support, security patches, or updates for these versions of our collaboration software will last through the below dates
Zimbra 9.0.0 End of General Support: 03/31/2024
Zimbra 8.8.15 End of General Support: 12/31/2023

[jeastman]

Hi oetiker,

Thank you for your feedback. No matter how critical someone is, it is always best to hear their opinion.

Would you mind explaining a bit more about your concerns?

The support lifecycle dates of the product versions have been adjusted in the past to account for a variety of situations. 8.8.15 has been extended (several times, if I am not mistaken) to accommodate many customers who felt they needed additional time. 8.8.15 was released in July of 2019 and is long overdue to be retired. Zimbra 9.0 was released in April of 2020 with an expected 3-year lifecycle. This too was extended to account for the release of Zimbra Daffodil (aka Zimbra 10), providing one year between the general availability of Zimbra Daffodil and the end of general support for Zimbra 9.0.

With the release of Zimbra Daffodil we have also focused on installation, upgrade and migration aspects of the product. In many cases it is easier to upgrade your Zimbra system to Zimbra Daffodil than versions in the past. We will be continuing with this effort to cover many more variations in the future.

I am interested to understand your concerns around these end of support dates and how we might be able to address those concerns.

Thanks.

[oetiker]

Hi

As I understood I can't upgrade with NG installed...

cheers Manuel

[jeastman]

Thank you for clarifying oetiker.

There are currently some issues we have seen with the upgrade of single-server systems where certain NG modules were used. We are suggesting that folks in that situation hold out until we have a resolution. Depending on your setup (which NG modules are in use) there may be options. We have some work-arounds for multi-server systems and are in the process of productizing those workarounds. This is a data portability problem and one of the reasons we have move away from some of the components used in previous versions. We strive to provide complete data transparency and want you to have full control of your data. When a component prevents you from accessing your own data, I see that as a problem and thus the change. Unfortunately, the very nature of this makes it difficult to move your data to a new system. We are actively working to provide a solution for this and will keep everyone updated.

This is a high priority items for us and something we will have a solution to. As I've stated, it is not right for a system to hold your data hostage.

If you can tell us more about the specific components you are using, we may be able to help you upgrade should you wish to do so.

[Klug]

As I said in another thread (about Daffodil) the "workaround" you propose to migrate data (multi-servers setup) out of NG-HSM to your HSM is very problematic (read: unusable for many customers).

On the new server we're suppose to provision enough space as primary storage for all emails. No deduplication and no HSM (secondary volume).
That's multiple terabytes of primary storage to provision.
Plus, of course, the previous primary/secondary storage (old server) and the new secondary storage (new server).

How do you suggest we deal with this?
For some customers, it means they cannot upgrade to Daffodil because they just don't have enough free space to do it (and they cannot add multi TB of storage before next year - if they can).
I'm thinking of public organizations (town halls, schools, universities).

Also I think it quite strange to suggest NG-HSM "holds your data hostage".
Until Daffodil, it was the official module and the supported way to deal with HSM (and I do think you (Synacor) were quite happy to find/adopt this solution in 2017).
Now you have your "own", it's a bad solution?

[L. Mark Stone]

I thought it would be helpful to add some context here for those who don't know John Eastman.

John joined Zimbra when Zimbra bought Mezeo in 2014. He started his career as a software developer and last summer was promoted to be the head of Zimbra Product and Engineering. During his tenure at Zimbra he has held several senior engineering/technical roles as well as client solution and business development roles. He knows Zimbra as a technical product; its history, and; he knows how different classes of customers leverage Zimbra from a business standpoint.

IMHO there have been improvements in Zimbra since the buyout and the resulting senior management changes. The online advertising and portal businesses have been sold, leaving Synacor to focus on its core businesses: CloudID and Zimbra. The new CEO of Zimbra in a public interview stated that a goal is to make each of Zimbra and CloudID as separate organizations as possible, empowering each new management team to make decisions that are optimal for their own businesses.

Having said that, John inherited the results of decisions/actions taken by the previous management team from when Zimbra was public. In my first career, I did technology mergers and acquisitions for 13 years, and later led a $5 million software development effort, so I could see that the new management team, and John in his new role, inherited some significant challenges based on those decisions/actions taken by the previous management team.

One of those challenges, which this thread discusses, is providing as seamless an upgrade path as practicable for existing OSE and NE customers who wish to migrate to Daffodil. For me it's quite personal as my company currently has no supported, seamless-to-customers upgrade path for our multi-tenant Zimbra NE-with-NG 8.8.15 hosting farm. And as others have already posted, time is tight.

John and an increasing number of Zimbra developers/engineers review and post here on the forums. The more the Zimbra team hears from OSE and NE customers regarding your concerns/needs/issues/RFEs etc., the better.

With best regards to all,
Mark

[jeastman]

Mark,
Thank you for the re-introduction. Much appreciated.

Klug,
Workarounds are just that. Something to overcome an obstacle until a durable solution is in place. The team is actively working on a solution to make the entire process more optimized and easier for people to execute. I'd love to say it all "just works", but the reality is that it is quite a challenge. I don't see this platform as a place to support marketing spin. This forum exists to address real problems with real technical solutions. Providing the best technical solution to the customers and users of Zimbra is what I see as the goal of the Zimbra team, so you will hopefully not see a lot of hand-waving and glossing over of issues here.
I have been pushing my team to get more involved in the forums here so that they can hear, first hand, the issues and concerns people are encountering. The hope is that this will help them to better understand the impact of their decisions as they continue development of the product.

Thank you for your feedback. The voicing of Your frustration, concerns, and challenges help to make Zimbra better. The issue you describe is a real one and something the team is actively working to solve. I am confident that with your help (and everyone else providing their own feedback) we can provide a solution that will work for all cases.

[Klug]

It's not frustration it's the feeling that "we're stuck".

We have to migrate away from 8.8.15 at the end of the year but we cannot, because of the storage (multi-servers), because of the modern UI, because migrating a single server to 10 is not yet possible (while we're in april and we all know the QA issues - see last patch).


Most customers are still running 8.8.15 (at least in this part of the world).
They did not upgrade to 9.0 because they could not find any interest to it, it's actually the opposite.
Modern UI is not getting adopted because it's new (and we know people don't really like new things) but also because it lacks key features of ZWC such as the minical (and adopting it is costly in training/time/support).
I'm quite sure you actually know this and that's why the Classic UI was kept in 9.0 and even in 10.

However, AFAIK, there's currently no way to disable the Modern UI in 9 (nor 10).
Users can switch to it (like they currently can switch to HTML view, that's more or less the same issue, it creates quite some support cases to get users back to the AJAX view).

There's a bug (ZBUG-1861) that says "resolved" since february 2022 but no "how to do it".
When ask, supports gives you the zmprov variable "zimbraPrefClientType" that contains the "user preference of client type" but no way to disable Modern UI (case 01318650 - I re-opened it yesterday).

So my current idea is that if we had a way to disable Modern UI (disallow users to switch to it, based on CoS/user), we could migrate to 9.0 and "win" 3 months of general support and one year of technical guidance.
That would allow everyone to prepare to migrate to 10 (customers to prepare the needed infrastructure and test Daffodil, Synacor to deal with QA/migration procedure, etc).

[thomas.klaube]

Hi all,

we support dozens Zimbra installations, most of them single server, some multi-server. 2 (that is "two") are running Zimbra 9.0, all others are on 8.8.15. The reason was already pointed out by Klug - our users are just not interested in the "modern UI" as it lacks soooo many things.

I have, at this time, absolutely no idea how we should migrate all these servers to Zimbra 10 by the end of this year... There seems to be no working migration tool or procedure. If I am missing something, please give me some hint....

Thanx and regards
Thomas

[rainer_d]

Hi,

we're also on 8.8.15 (latest patch minus 1), on CentOS7.

I've basically just migrated everything to S3 backup and HSM because I - foolishly one might say - assumed it would be much easier to do migrations once you don't have 2TB of data sitting in a VMWare volume. The documentation even pointed that out (in some way)...

Maybe an option would be to EOL Zimbra 8 on platforms that are EOL anyway (CentOS 6+7, old Ubuntu releases) and then continue to build 8.8.15 patch releases only on recent OSs.

Of course, we don't know how difficult it is to maintain 8.8 and where the "pain-points" with it are - AFAIK, it has never been communicated. From cursory looks, it seems a lot of components were updated over the years (MariaDB being an exception).

I admit I haven't spent too much time on upgrading even to 9. I have other duties and I just "chase" the patch-releases - it's planned for Summer, though.

If you (Zimbra) think you can figure out something in the course of the year, that's good enough. My plan was to upgrade to 9 through Summer/Autumn and then I hope there is a sensible path to 10 without undoing all the S3-stuff I have (and which took a rather long time to be actually usable).

I don't really want to put any more pressure on you guys, but please: Help us