Create
cancel
Showing results for 
Search instead for 
Did you mean: 
Sign up Log in

It's not the same without you

Join the community to find out what other Atlassian users are discussing, debating and creating.

Atlassian Community Hero Image Collage

Answering your questions on changes to server and Data Center Edited

Bryan Mayo Atlassian Team Oct 16, 2020

Hi Atlassian Community,

My name is Bryan Mayo and I’m the Head of Customer Support & Success at Atlassian.

As you might have seen in our announcement blog, we recently announced some changes to our server and Data Center offerings to sharpen our focus as a cloud-first company and deliver the world-class cloud experience you deserve.

We know that change of any type can be hard, and our teams at Atlassian have been working hard to make sure you have the resources you need to navigate these changes and plan for the future. You can visit our website to access these resources and get a summary of what’s changing, a detailed timeline, answers to common questions, and recommended next steps.

While we did our best to proactively answer your questions, we know there are some we missed. If you have any additional questions about these changes, please leave them below. We have a team of Atlassians standing by who are ready to help.

68 answers

This is an incredibly bad idea.

I've been an Atlassian administrator for several years now. I have worked with a lot of companies, big and small. Some on Server, some on DC, some on Cloud. Many, many companies do not want Jira Cloud. They need their data on a server they control, behind a firewall they manage, accessible without the internet. They have customizations that won't work on Cloud. They rely on Add-Ons that aren't available for Cloud. And many of them are companies that can't justify $40k/yr for Jira alone.

Killing off Server is an incredibly backwards, harmful move. I see no technical reason to do this, it seems solely to force people onto Cloud or Data Center, where you make more money. This move seems entirely driven by greed and profits, not what's best for the customer. Keeping Server alive, as a product, has minimal overhead for you as a company. It's basically the same code as Data Center, which you aren't killing (yet). So it's not like Server was losing you money by existing. The only reason to kill it is to FORCE people into things that make you more money. That's bad for the customer.

Before, a customer could try Jira On-Prem for as little as a $10 starter license. Now the cheapest option they have is, what, over $40k? That is insane.

This is going to leave a huge gap in the market that some other company is going to come along and fill, stealing all of those customers away from you. There are a lot of companies out there with under 200 employees that have no interest in Cloud and can not afford or justify the cost of Jira Data Center.

https://www.atlassian.com/company/values

"Don’t #@!% the customer"

Hey Taylor,
I hear you - this is a monumental change, especially for our server admins. We’re deliberately investing in programs that ensure that you are set up for success for this transition - Cameron Deatsch, our CRO, and former Head of Server and Data Center captured these plans in his post To all server champions we want to hear from you. Read that post and considering applying to join this ongoing initiative.

Respectfully,
Stephen Sifers

Data Center sits behind the firewall and offers low to no-downtime upgrades.

Like # people like this

Hey Stephen,

I tried to reply to the thread you linked, and seem to be blocked on doing so. 

Taylor

I have to agree with Taylor. It is an incredibly bad idea. I have read the Security and Compliance portion of "Debunking the Myths" and find it misleading at best. All of the "answers" are hearsay at best.

It claims that "typical" installations have one login that is shared. But I know that many don't. So that argument doesn't apply.

It claims that techs use "Shadow IT." In a secure environment, techs cannot bring shadow tools into the environment. Systems are monitored so that if someone DOES manage to bring one in, people are notified that a new program is running. If a tool is needed, it can be tested and approved, or, if it represents a security risk, the employee is notified.

No, the real fact is that even some security vendors do not understand security. Many believe that a firewall is a device designed to block things. Not true. A firewall is a device to GRANT limited access. 

The "cloud" by its very nature is less secure than on-prem.

Not everyone is able to provide the same level of security that Atlassian cloud does, but many are able to provide even more security. The move to kill server prevents those that have that ability from keeping their data secure.

I am deeply disappointed in Atlassian for this move and will be looking for alternatives to recommend.

Like # people like this
Bridget Community Manager Oct 16, 2020

Thanks for this post. Perhaps you've already digested the materials in our Cloud Trust center, but here is the link in case this site can provide you with any more information/assurances. 

I also urge you to sign up to be part of our server champion community group so that you can give this feedback to the team. You can find signups and some more details about what we're doing to support our server champions through this transition here.

I hope these resources are helpful, and we will document your above feedback to the team. Sincerely,

Bridget 

Like Martin Heyer likes this

Bridget,

Thank you for your info. I do not have my name on here or the name of any company I work with because of NDA agreements. However, I can tell you that in least a couple of environments, the development systems CANNOT access the Internet for ANY reason. Anything that needs to copied the dev environment must go through gateways systems that do not have direct access to the Internet.

A cloud version of Atlassian products would force a hole into the development environment, compromising security. You can have multi-factor, encryption and all kinds of security patches on top of that hole, but it is still a hole. And the company is not large enough to afford or need cloud versions. 

However you try to paint it, it is still a grab for more money and, especially with the losses that have been felt because of COVID-19, it feels like you are trying to bite the hands that feed you in order to get more. Do not be surprised if those hands pull back and stop offering food.

Like # people like this

We have no interest in Cloud.  In fact we were originally on Cloud and went to great pains to migrate to Server.  Never been happier.

Cloud, in spite of the rhetoric, is a totally different product (somewhat dumbed-down to put it crudely) and will not serve the needs of many businesses.

This is an absurd move on Atlassian's part.

Like # people like this

Worst. Idea. Ever.

 

"We know that change of any type can be hard, "  This change is much more unwanted than hard! Not everyone wants to move in the cloud. I really don't want to give up my server in any case and i don't want to pay for a much bigger license i need. Not everyone really "needs" a cloud. Many users are more than happy with a "normal server".

 

Can't tell in words how disappointed i am.

Like # people like this

How can you state you are GDPR compliant on https://www.atlassian.com/trust, while basic functionality is not. Some of the Critical GDPR / privacy bugs are open since 2004, and you don't care about them at all.

 

E.g. 

https://jira.atlassian.com/browse/CONFSERVER-1882

https://jira.atlassian.com/browse/CONFSERVER-7837

 

https://jira.atlassian.com/browse/JRACLOUD-36896

https://jira.atlassian.com/browse/JRACLOUD-7776

Like # people like this

I am 100% in agreement with every word of your message @Taylor Huston .
Atlassian, stop thinking about money and think about customers. This change is insane, simply unfeasible for many clients! I hope they reconsider

Like # people like this

This is horrible!

We invested money and time on Jira and Confluence to setup our documentation and procedures using add-ons that are not supported on the cloud version. We also have other requirements that make the cloud version not suitable.

So, the only option left is to buy a data center version but the price difference is gigantic and won't happen.

To me, it's simply a decision of getting rid of "complicated" customers. Your message is heard loud and clear!

Like # people like this

Our confluence servers are integrated to other systems and databases and we will not change to a cloud based solution. So we have four years to find another product and for a migration. If this is really not a very early April joke, we'll say good bye Atlassian...

Like # people like this

Time to leave Atlassian. Any recommendation for alternative issue tracking software that is customer friendly? Dig up good 'ol Bugzilla again?

Like # people like this

It appears that our relationship with Atlassian now has an expiration date.

We will not convert to an off-premise solution for the reasons many have mentioned above.  

A company that would strategically choose to take this route is not one that I will continue to waste my time implementing. 

Like # people like this

We have integrated confluence in an complex network environment with databases and other services. We have no interested to open this network to the atlassian cloud. As university institutes with 250er license we simply cannot pay the datacenter license, we already pay 10.000$ per year (and started with 3.000$!!!). So also the last price increases had been heavy for us. We stayed simply because it was more expensive to migrate. Now it is clear - we have to develop a tool moving from confluence to some other solution which is more respectful with their customers.  Good bye Atlassian

Like # people like this
33 votes

Best of luck Atlassian but this is not a journey I am interested in. 

To be completely honest, this feels like a complete revenue raising initiative that is going to burn bridges with existing customers.

I have implemented confluence server in every job I have been in for the better part of the last 10 years and have always had ongoing maintenance for everyone of them. I love the product but not enough to fork out the exorbitant fee of Data Center. I could get it if you just increased the price of Server (by an acceptable amount) but this is just crazy.

I have absolutely zero interest in cloud and in some instances cloud is not ever an option due to internal requirements to keep the sensitive data on prem.

 

At least I have a year to find an alternative to Atlassian products and migrate off then decommission the existing servers we have.

I want to use substantially more colourful language to describe my feelings towards this but that would not be appropriate or helpful.



Bridget Community Manager Oct 16, 2020

Thank you for the constructive comment. Our ultimate goal here is to unleash the potential of every team, and this shifting of investments is simply to support where most teams want to be today and in the future. 

That said, we really empathize with your frustration here, it is a huge change that (of course) Atlassian put a ton of thought and care into, but that still doesn't make it easy to hear for our longtime server customers. 

We know it take time to figure out the right decision for you and your company, and in the meantime we are doing everything we can to support server admins like you. If you haven't already, read out this post and please consider signing up to join our server champion community group. 

Sincerely,

Bridget

Like Gülçin Nacak likes this

Bridget, I think a lot of the frustration comes from the realization that this change wasn't done to help us, it was to keep your support costs lower and increase Atlassian's value to shareholders. Forcing a move to DC just to get support past 2024 (while increasing the cost of both Server and DC at the same time!) just doesn't sit right when Jira Cloud isn't as fully-featured. Is the code base so significantly different that Atlassian couldn't offer Server and DC? If it's not, then this is simply an attempt to force SMBs to a less-functional cloud product. So many companies have been hit very hard by the pandemic, and increasing costs now, of all times, feels...wrong.

Like # people like this

> Our ultimate goal here is to unleash the potential of every team,
> and this shifting of investments is simply to support where most
> teams want to be today and in the future. 

a quite selfless decision from Atlassian ...

Like # people like this

Forcing a move to DC just to get support until 2024

I'd like to notice that Server is supported till 2024, the is no EOL for Data Center, so having an on-premise installation (using DC license) with maintenance and support beyond 2024 is absolutely possible. 

Like # people like this

Marek, I think everyone here is aware that we could move to DC if the cloud offering cannot meet our regulatory / security / privacy needs.  Two problems though:

1. Compare the pricing for a SME with say 20 Server users currently versus the minimum tier of 500 users for DC.  Small companies simply can't afford that.

2. Why would anyone trust Atlassian ever again?  Who would be naive enough to think that they're not going to get another cheery email in a year or two's time saying that DC is being dumped?

I understand that we are the most difficult customers.  Big companies have full IT teams and even develop their own add-ins.  Cloud customers are cookie-cutter, lowest common denominator accounts.  Cheap to have.  We're the ones who test your product development teams, and people like me who have been an customer and admin for 17 years are the ones who grow into bigger businesses and recommend you.

So I think this is a very short-sighted decision.

Best of luck.

Like # people like this

> 2. Why would anyone trust Atlassian ever again? Who would be naive enough to think > that they're not going to get another cheery email in a year or two's time saying that
> DC is being dumped?

or maybe the price for DC will be increased to 100.000 or 200.000 USD a year. Who knows? That's really a matter of trust. Beside of the fact that we are not going to pay for DC version as a small company.

Like # people like this
27 votes

How can you expect your customers to be ok with the HUGE price increases that will come from this ... especially during these times?

For instance this is what we pay annually for our Atlassian Products and apps.

Server - Total $38,643
Confluence -  $33,298
Jira Software - $5,345

To move away from Server to Data Center and retain all our current functionality would be ...

Data Center - $97,968
Confluence - $74,496
Jira Software - $23,472

To move to Cloud and lose much of our functionality as the apps don't exist

Cloud - $137,640
Confluence - $128,840
Jira Software - $8,800

Just to move to Data Center it would cost us over 2.5 times as much. There is no way that our organization will go ahead with that. They will tell me to start shopping around and planning a migration. Given that how does this fall in line with your value of "Don’t #@!% the customer"? Please, I love your products, but you are seriously missing the mark on not "#@!% the customer".

Yeah, this is not great. We use add-ons that are absolutely not available in the Cloud version. If we have to change our workflow, we might as well look at other products.

Like # people like this

I just moved us from Cloud to Server because the add-ons were not available in Cloud.  The add-ons allowed us to be 21 CFR Part 11 compliant.  I can't go back to cloud unless that offering seriously ups its game, and I can't afford to go to Data Centre.  Unless something changes it seems I'll have no choice but to move elsewhere.   Sharepoint maybe.

How about splitting down the DC 1-500 user pricing into smaller user number categories so that it accommodates the SME companies who are currently using Server?

Like # people like this

Agree, Geoff. We have a 100-user seat and are back at 49 active users. The price increase to DC is insane.

Like # people like this

Agree. In many respects server is just a better, more functional product. Moving this expense to opex is just not an option for us. Especially not considering we would pay double or more or weaker functionality. 

Like # people like this
Emilee Community Manager Oct 19, 2020

@Stephen Reece @Geoff Daly _Admin_  What are the server apps that you use that are not available in cloud? I'd be interested in understand what apps you are using that don't have any cloud options or alternatives. 

Like Stephen Sifers likes this

It's not the the apps. Some is just configuration

  1. Custom HTML - I want to add a bit of jQuery to the page to tweak some details
  2. Custom Stylesheets - I want to hide buttons from remote users and make the styling different, especially for common macros
  3. The ability to hide internal comments from anonymous visitors!
  4. Custom URL - makes a HUGE difference in SEO with Google. The cloud version is constantly screwing up our search results
  5. Global Search/Replace plugin is much better on server

Are you going to support these in cloud? If yes, when? If no, can you spin off the server product into its own company?

Like # people like this

We use "Jeditor" and "Links Hierarchy" on Jira Server that do not show cloud versions.  We also use database tools like BIRT and ODBC that are not configured for cloud use.

Like # people like this

Some examples mandatory for us:

* Confluence: "Space privacy - extranet" being able to see/mention ONLY users with permission in your spaces

* JIRA: Tempo Budgets . No alternative with all features

Like # people like this

@Emilee our Salesforce CRM connector add-on is one of the biggest. along with Epic Roadmap and Workflow Enhancer. There may be Cloud add-ons that approximate it, but there are limitations that add-on devs face with Cloud that they do not face with on-prem.

Like Taylor Huston likes this

@Stephen what about custom workflows in Jira and a good look and feel which really works.

Davin Studer Community Leader Oct 20, 2020

Confiforms is a big one that we use and while there is a Cloud version it is more limited and doesn't work at all with the new Fabric editor since you messed up nested macros.

Feeling a bit betrayed on this one. As a nonprofit, we have to keep costs down, so the server option made this solution possible. We are also technology centric, so hosting ourselves is not a problem. We're barely months into implementation, and now this. This pricing change quadruples our cost. Are there going to be academic/public library discounts in the new pricing model?

Atlassian will definitely lose customers whose security policy prevents them from keeping this kind of data in the cloud, as well. While I can sympathize with the difficulties of developing for both hosted on on-prem systems, these types of systems really need to have both options.

At the very least I did find reference to some plans for nonprofit and academic discounts. That will be helpful. Since we are still in implementation (albeit deeply), we can use the free cloud migration as a way to evaluate whether the cloud version will work for us or if we should look elsewhere.

Maggie Atlassian Team Oct 16, 2020

Hi Sean, 

Thanks for the post. Yes, your server discount will still apply if you continue to remain on your server license (and you can continue to do so for three years). Additionally, it sounds like you found our academic and community licenses for cloud, but I'll share that resource with you as well in case it is helpful. 

Let me know if I can help with anything else.

Like # people like this

Same issue here... what will be the non-cloud options for non-profit organisations?

Same here.
We have a community license for non-profit organisations.
@Maggie, will we pay after february or our discount still apply until february 2024 (apps include) ?

Hi.

That is the absolutely worst what you could have done to me any my company. We have built everything on Atlassians products (Confluence, Jira, Bitbucket, Bamboo) and we have trusted that you would not stab your customers in the back.

We can't migrated to the cloud and we will never be able to do so. Also as company of 100 ppl we wont be able to pay DC. You left us without any option.

No idea what to do.

Regards

Yep, it's utter nonsense. Can't wait for their stock to tank.

Like # people like this

Sorry, I guess I accidentally deleted my last post when trying to edit it (?)

I have a question - looking at this page, it looks like the majority of updates to the server offerings will end on 2/2/22 along with all updates to end on 2/2/24. After 2/2/24, will we still be able to use our on-prem server products (in my case, Confluence and Jira Service Desk) but updates and fixes will cease to continue?

As for my own feedback, no one likes price increases - especially very steep ones, and no one likes to be told that the product they've used for years is going to stop being supported. But I am thankful that you've given us a few years to get things figured out, we've seen worse from other companies. There was no mistake for the reason as to why this news was released on a Friday afternoon/evening in the US (making it very late Friday night to early Saturday for the rest of the world's majority). But it looks like the announcement did good things to your stocks so mission accomplished on that, right?

But the truly insulting part of this is the continued insistence that we know nothing about "the mysterious cloud" and why we should just go to it since Atlassian thinks they know better than us. I would double check with your marketing team to make sure they know it's not 2015 any longer. The whole "why you shouldn't be afraid of the cloud" tactic is becoming antiquated and anyone in charge of a large number of (potential) licenses knows the risks, the advantages, and the disadvantages of moving to the cloud by now.

In 2020, especially well into the COVID-era, you are already in the cloud, in the process of moving to the cloud, or staying with your on-prem server for a very good and calculated reason. This whole thing isn't going to force us to store our crucial and confidential data on your cloud, it's going to force us to either continue using your unpatched software on an ever increasingly secured server (which sits right next to me) or will make me look at another product to go with.

We're not afraid of the cloud, we use the cloud already with nearly 100% of what we do in our company. I appreciate that you offer your products on the cloud for those which works best for them. But your cloud offerings fall short in so many ways important to us, compared to the server services. I'll leave it with this: if you didn't offer your server software in the first place (and only had the cloud offerings), we wouldn't have gone with your products at all due to their shortcomings.

(New account created for this post as I choose not to have my real identity revealed)

@Tom Shaffer - to answer your question: Yes, for server you bought a lifetime license. So you can choose to run your Confluence and JSD server as long as you want to.

This also includes all the marketplace apps you may have.

Like # people like this

Thank you, Matthias. That will help us (and I'm sure others) know all of the options they have, even if it may not be the most ideal one.

Like Matthias Gaiser _K15t_ likes this

we migrated our confluence server version to docker, so we can keep the last version als long as we want - up to now we have a open link to the internet, so we will close this after the last security update. This will save as time to move to a new version. Hopefully there is a market for somebody offering migration tools from confluence to something else in the next years. For security reasons we will never move to the cloud with our vital data. In fact, from europe we are simply not allowed to move to servers not in europe.

Like YannickGiguere likes this

Terrible idea.

What about installations that require data and documents to be kept locally and not not on some server somewhere on foreign soil?

They don't currently support data localisation.

Like Taylor Huston likes this
Maggie Atlassian Team Oct 16, 2020

Thanks for the question. We will continue to offer Data Center, our self-managed enterprise edition for those of you who need to maintain control of your data, and we will offer loyalty discounts. I'd also encourage you to continue to check out our cloud roadmap to see what is upcoming as we continue to work towards meeting more of your needs in cloud.

Data centre lowest tier is 500 users.

What about small to medium sized businesses with the same localisation issues?

Like # people like this

The loyalty discount  is a joke for small teams.

Like # people like this

Small teams is exactly how I began using Atlassian products. If the server license had not existed eight years ago I would never be a customer now.

Like # people like this

So, you are saying there is only a moderate price increase, but as I see it our 10-user Confluence license is increasing from 10$/yr to 1300$/yr? That's a 13000% increase?

Also, do you cloud products already support trivial things like 'hosting on your own domain'?

In 2018, we purchased numerous products (academic discount and 36 month term) that worked out to under 20K a year. The same products now? 50K. 

We've built out workflows that (I hear) won't translate into the cloud. 

That we are compelled to move the the cloud is disappointing but not at all surprising. The price increases? Products we looked at 3-4 years ago are now worth another look -- esp. since the transition to the cloud is likely to be so disruptive for us we might easily rebuild in another product.

Like # people like this

Which products you are looking into now?

Like # people like this

I have multiple connections to Atlassian, and I'd like to give you some insight from each of my perspectives about what this news means to me.

 

(1) AS AN IT CONSULTANT AND AGILE COACH

I have been an advocate of Jira and Confluence for many years. I've proposed and supported the introduction of Jira and Confluence in large companies. All of these companies would never consider a cloud product because of regulatory rules, legal provisions, and/or internal standards. You're obviously forgetting that there are a lot of good reasons why companies opt not to use cloud services. 

In those companies, I would not have been able to convince anyone to try a cloud solution.

You might argue that you are still offering Data Center, but you've made your path clear: considering your marketing, and now drop of Server, and these extreme price increases for the Data Center solution, the trust is gone in that you will continue to support hosted solutions in the long run.

 

(2) AS A CUSTOMER

While we advise large companies, we ourselves are a small company with less than 10 people. For our hosted solution, the price increase is 1600%. And we're not sure that is the end. What stops you from increasing prices again and again?

The price policy reminds me of IBM's when, in the 1980s, they started to squeeze out their mainframe customers by continuously increasing prices, trying to disguise it through repeated changes of how the prices were calculated. The result was that their customers fled from that platform, investing a lot of money and manpower, just to migrate to more calculable platforms.

 

(3) AS AN APP DEVELOPER

We have developed apps for Jira and Confluence Server which provide functionality that our (and your) customers were missing. We have invested a lot of time and money in these apps. Unfortunately, it is not possible to provide these apps for Cloud because of the limitations there. Cloud apps depend much more on an integration option from you, Atlassian, because the environment is much more limited. But although we were asking for hooks to integrate our apps, that was never done (and probably never seriously considered).

We know that other app developers have the same impression, it's not only us.

So for us, the news means very bad news. We expect most of our customers to go away within the next 1-2 years. So it does not make sense for us to continue investing in app development. The Atlassian ecosphere will definitely become less diverse.

 

(4) AS A SHAREHOLDER

I became a shareholder of Atlassian not because of the unmatched product features or quality. Rather was I impressed by Atlassian's sophisticated licensing model. With the support of small installations, I saw that many Jira and Confluence advocates in large enterprises were actually users who had tried out Jira Server in a small environment (even at home!). Many consulting companies (like us) are small companies, and they were using the same products which they recommended their customers.

While it was totally clear that you would not make money from these small companies, I was impressed that this was actually an extremely intelligent way of marketing!

However, either I have overrated Atlassian's smartness, or this has been forgotten about in the last cost effectiveness reviews, and requested to be dumped by a financial controller who does not know about how this market works.

Your prices increases have caused Atlassian shares to rise significantly over the weekend. This seems like good news for shareholders. Maybe for the next 1-2 years, maybe 3-4, but I'm so convinced that these decisions are wrong in the longer perspective that I now have a pending sell order.

 

Please excuse me for not using my real name for this post.

Me too, I always thought $10 license is smart, use it when you get big we will take it back when you can't leave anymore and it happened with us, jumping from 10 users to 25, but they lost someone that was smart.

Like Taylor Huston likes this

I am a home user. I have repositories that I do not want stored in the cloud. I want them on my own server, that I can be sure is not being accessed by anyone else, and that I know will be accessible 24x7 regardless of any Internet outages. I have a Bitbucket Server Starter License, for which I pay $15 a year, and I've been an Atlassian customer for 8 years. From the pricing tables it looks like I will now be billed $1650 a year. So a 110x price increase. ONE HUNDRED AND TEN TIMES, or ELEVEN THOUSAND PERCENT. Seriously. I am a home user, I am retired and I don't have that kind of money. Way to go Atlassian. I'll be moving to Github or Gitlab

It is now 3 days later, and I have moved my repos to my own Gitlab server. Was it painless? well in all honesty no, but there's a lot of useful help out there and it's just a case of resolving each issue as you go.

My server is running Debian buster. It's an internet facing machine and has other things running on it (Jenkins, Bitbucket, Apache - including a number of Rails sites). This is probably the first hurdle. If you were starting with a clean machine then things would be easier. I followed Gitlab's instructions to install a gitlab-ee package. There are some pre-reqs like postfix and ca-certifcates which I did first. The instructions explain this. Running the server did not immediately work and that's probably because my install didn't complete, which is probably due to me already having some LetsEncrypt certificates and already having port 80 and port 443 in use. So I followed lestEncrypts own instructions for getting a new cert (more on this later). There are also issues with setting up the service to start Gitlab and getting it to run on boot (maybe due to my partial install).

Next I wanted to not use a local PostgreSql but my own server which is not internet facing. That's not too hard though, edit Gitlab's rb config file and reconfigure. I created an empty database and Gitlab automatically migrated it. I also wanted to use Apache not nginx and again that's just a config rb change and a bit of Apache set up - tricky bit is the port redirecting). My vhost config may help some:

<VirtualHost *:<myport>
ServerName gitlab.<mydomain>
DocumentRoot /opt/gitlab/embedded/service/gitlab-rails/public

ErrorLog ${APACHE_LOG_DIR}/error.log
CustomLog ${APACHE_LOG_DIR}/access.log combined

ProxyPreserveHost On
AllowEncodedSlashes Off

<Location />
Order deny,allow
Allow from all
Require all granted
ProxyPassReverse http://127.0.0.1:8080
ProxyPassReverse https://<mydomain>/
</Location>

RewriteEngine on
RewriteCond %{DOCUMENT_ROOT}/%{REQUEST_FILENAME} !-f
RewriteRule .* http://127.0.0.1:8080%{REQUEST_URI} [P,QSA]

RequestHeader set X-Forwarded-Proto "https"
RequestHeader set X-Forwarded-Ssl on

# Cert location
SSLEngine on

SSLCertificateFile /etc/letsencrypt/live<mydomain>/fullchain.pem
SSLCertificateKeyFile /etc/letsencrypt/live/<mydomain>/privkey.pem

Include /etc/letsencrypt/options-ssl-apache.conf

</VirtualHost>

I've replaced my real domain name and real port number here. I also needed to enable the request_header module. Because I had already created my letsencrypt cert, it was in the folder listed here (/etc/letsencrypt ...)

That was pretty much it

Like # people like this

Next I need to move 65 repos over. That's pretty painless but quickly threw up another issue. By default Gitlab stores its data in /var/opt/gitlab. On my machine /var is small - I only have log files there. I needed another config rb change to set my Gitlab data folder to /home/gitlab/git-data (I just moved /opt/var/gitlab/git-data to /home/gitlab where I have a load of space). Before this I was constantly filling my disk and causing upset everywhere.

Migrating repos is easy, but remember a repository in Bitbucket is a Project in Gitlab. The best metaphor for a Bitbucket project is a Gitlab Group. I used projects in Bitbucket (e.g ruby, python, node etc for the type of repository) I created groups in Gitlab. Then I tried to migrate my projects. There's a mechanism for importing projects directly from a Bitbucket Server and I started off using that. It needed a bit off setting up - creating personal access tokens, but that's quite straightforward. However for some reason the import page only showed me about 20 projects. I imported these OK, but maybe it was because I was filling my disk up, I couldn't automatically import the other 40-odd.

So Importing one project at a time was the way to go. It's not hard, just give the url for the .git file in it's Bitbucket location and import it. Look out for not getting your credentials right though - the import will go quiet on you and will take 3 hours to time out - or you delete the project and try again. I did this yesterday evening while there was nothing on TV - it's a mindless repetitive task.

Like # people like this

Then today I updated the repos on my dev machine. I'm still using SourceTree, but if Atlassian stop supporting that, I'll look for another git IDE - sorry I can do the command line, but why complicate things! Here it's again straightforward and my one stumbling block was ssh. I needed to add a rule on our home router to NAT the port I would using in my git urls (e.g git@mydomain>:<an ssh port>/<project>/repo.git if you can make sense of that. I have not opened up port 22 to the outside world, and I have locked up the port I use so that an rsa key must be used. It's not hard, just a setting in /etc/ssh/sshd_config and make sure your id_rsa.pub is imported in Gitlab. The sshd_config setting is:

#GitLab
AuthorizedKeysCommand /opt/gitlab/embedded/service/gitlab-shell/bin/gitlab-shell-authorized-keys-check git %u %k
AuthorizedKeysCommandUser git

It basically tells ssh to run gitlab-shell-author... to check keys.

Now I edited the settings for all my repositories through Sourctree. Pretty mindless, open each repo, select repository setting edit the "origin" remote path to look like the git url I showed a few line above here. Change the Host type to Gitlab CE and add my own user name. Then I tried a fetch to verify all was good. Oh, to get the url, I opened up the project in Gitlab's web page and copied the clone url to the clipboard - in some cases I have renamed the repository, but in others I just edited the url, changing my sub domain from the bitbucket one to the gitlab one and changing my port (remember this is the ssh port not an https port)

Like # people like this

So there you go. I said it's been 3 days, but I've been out for a walk most of today and in reality it turned out to be just about 6 hours of work and most of that was figuring out how to get over the hurdles I'd put in place myself (a server with other stuff on it, a different Postgresql Server, Apache rather than nginx and of course not having a big enough disk).

I still have some work to do. My Jenkins configs will need to be changed so they use the Gitlab server rather than Bitbucket and I will have to change settings on my Rails sites so Capistrano deploys from Gitlab. Again, not too hard. I may miss a few settings, but of the 65 repos I have perhaps half are projects I have not worked on for years.

Oh, and one big plus that I can see: Gitlab is a standard Linux apt package. It's updateable just like other packages. There's just one config file, which I have kept a copy of. Bitbucket used to be half a day's work to update.

I'm still following this thread, so if anyone has issues, I will try to help (assuming I can actually remember how I got round every problem!)

Cheers

Derek

Like # people like this

I am currently using Cloud and migrating to Server. Even with this news, I am going to continue. Here is why:

  1. Custom HTML
  2. Custom Stylesheets
  3. The ability to hide internal comments from anonymous visitors
  4. Custom URL - makes a HUGE difference in SEO with Google.
  5. Global Search/Replace plugin is much better on server

So we will stick with server "as-is" until cloud (or another company) supports these things.

I understand the decision, but the cloud version sucks right now.

Yeah. Been there too. Server is just a more functional product even before you get into the massively powerful addons that are only supported on server.

Like # people like this

I would love to hear the answer to the question: What are you (atlassian) going to do about the need to customize cloud? is it planned? When?

Personally, I think they would have been smart to spin off the server product into its own company.

We did a lot of effort to bring business processes finally to own hosted JIRA instance, as this is required in our case by regulator. 

We will immediately start searching for alternatives and if we need to split up again into 10 different tools, but still cheaper, we will go for it. 

Atlassian, you try to make yourself Apple, but you forget there are other great tools out in the world and as fast you think you can change the behaviour of your customers to be pushed into cloud or high costs for less risk (and trust me, big companies e.g. financial sector are not allowed to. Did you also ever heard about GDPR in EU) you will there will be other tools rising up and taking this customers. Maybe this is your plan, but we are depending on your support, but this is no support, this is a joke.

After five minutes googling I can find a lot good, open source or cheap alternatives. 

Making yourself to premium brand was not necessary, many people trust in your expertise and tools. 

Cloud does not solve everything! You will see!

 

bye!

This might be a good idea to set up a page how to migrate to open source solutions. 
I certainly would be interested. 

Edit: To be clear, this is not how I would like to see this go, but you are not giving any alternatives.

Like # people like this

Please see the replies I made to my own comment (a few comments above this) I've just moved about 60 repositories to my own Gitlab installation. It was quite straightforward. Just be methodical

Derek

Thanks for documenting your work @derekk19 . Unfortunately moving repos is the least of my worries. I think I'm going to have to write a custom solution to export a decade worth of Jira issues + comments +  attachments + custom integrations with addons. 

11 votes

This is disappointing news and my company for one will be forced to go elsewhere after many years of happily using Atlassian products. It's a shame that for corporate reasons you've chosen to discontinue the products you were providing us, but I understand it, that's business, if maintaining Server is no longer profitable for you then you need to stop maintaining it.

What sticks in the craw is the way that instead of telling us how sorry you are and wishing us luck in our transitions you instead insist on patronising us by telling us that you know our businesses better than we do and that we actually want to stop using Server, that we want to move to the cloud, that the cloud is the right place for usNo! Don't add insult to injury by telling us what's right for our businesses. Don't tell us that you're "Accelerating our journey to the cloud, together", because we never had a journey to the cloud. We had Server and Server was what we wanted.

While I understand the move to cloud based services ($$$) and less support issues. 

For a “minor” portion of your customers cloud services are simply not an option. It can be IP / lack of certification / regulations 21 CFR Part 11 or GDPR / company policy. Our Notified Body simply will not except external sources which we have no control over. Which puts our certification at risk.

Just the continuous upgrade path of cloud is a dealbreaker for the situation I am in. There is no way to check for data integrity of the contents. Filed a ticket for that but nobody was interested. 

Datacenter at its current form is also a no-go for anyone under 500 users. Just the pricing will not pass management review. Leaving mid size business in the cold. 

Any users that can work in the cloud probably already moved to it already. So either you are considering your remaining customers as dead wood that needs to be pruned for profits or someone is sleeping at the wheel. 

At least I will have some time to look at alternatives. Sad to have to do this after using your products for over 8 years. 

edit:For a more in-depth explanation : here

It is not a minor portion it is a huge chunk.

Like # people like this

I'll echo the sentiment of many here by saying that this is really sad news for our 50 user setup to hear. Obviously we've got three years, but if there is no climbdown for mid-sized customers, we'll be forced to go elsewhere.

The biggest issue is Bamboo. I think we'd be fine to clear Confluence and possibly even JIRA internally for a cloud move, but giving a cloud version of Bamboo access to our databases, application servers, all of our code and then having to shunt all of that data up and down all day just isn't going to fly. Why would we develop on-prem applications using on-prem databases with a cloud-based CI/CD tool?

So we'll look for alternatives. And with Bamboo moved there's little reason to keep using JIRA, Confluence, BitBucket and the rest. We support a lot of Open Source tools, and it's been a credit to Atlassian that we've been happy to pay for their software because of how good it has been. But we move and adapt, because that's technology. So we'll do it again, and move somewhere else.

And paying for data centre just to avoid moving? Not a chance. I'm surprised anyone below 100 users is even considering it considering how much extra that's going to cost. But then again, what a bind they've put you in...

It's a shame to see that Atlassian has taken this wrong path for its customers.

My personal opinion, I want my money in my own pocket and I want my data on my own storage.
Now you are welcome to name me old school, but I heartly don't care if the cloud is the future for anyone.

Why we don't use cloud products in our company is very simple:
.) We are an authority and not allowed to give out our data.
.) Especially not if the data is stored outside of our country, or if there is just a suspicion that this can happen.
.) We only use Atlassian products for our internal purposes
.) We have no branches that, would have to access data from a head office via the Internet.
.) Regardless, of whether it is legally possible, we will most certainly not store our system documentation and tickets anywhere in the world.
.) So why should I save my own data anywhere in the world?
.) Why should I load my data over a slow internet connection when I have my servers in the house?
.) Why should I put extra load on my internet connection?
.) Why should you pay more for a worse solution?
.) And why the hell should I pay monthly fees, to be allowed, to access my own data?

The cloud may be funny for you, but the cloud just doesn't make sense for us.

To offer DC licenses that are so ridiculously expensive as an alternative, is simply a mockery.
And we're only talking about software that doesn't depict our core business, but rather serves as documentation.

If your customers are important to you, you continue to offer an affordable on-premise solution for the broader SME sector.
Offering data centers with a small number of licenses at the usual server license prices, has already been suggested by some.

I honestly hope, that Atlassian will find the spirit, that has been spoken of by many and that it will follow suit in its strategy.
You still have about 3 months to do it.

Hi,

I agree with the majority of the replies here, that killing the server products seems to be an awful idea. In the first moment I thought it is my specific use case that lets me feel like this. However, the responses here are showin me, that I am not alone.

Is there any option that you can re-think your decision?

Are there any ideas how to give your customers more control over their data and where it is placed?

Is there any option to get a similar functionality at an about the same level of annual costs when the server products are going away?

Best regards,

Joachim

Hi there,

I think the solutions to many complaints is very simple: 

add smaller user tiers to DC with a reasonable price

 

To further sharpen your products and make the price more reasonable for small companies, you should add a restriction to the DC setup. E.g. for a licence size <500 users no multi-node-instance is allowed, everything must be a single-node instance (which is exactly what a server instance is more or less). I don't think there are a lot of use-cases where a company with 50 users needs a 8 node setup. There is not much traffic on these instance and a single node is enough. I think adding such a restriction inside the DC code could also be simple.

 

So for small companies / use cases you only have a "one node server" setup, for larger companies (>500 users) you could have a multi-node setup. This makes sense since these companies automatically have more traffic on their instances and need the scalability of a multi-node DC.

This is a fantastic idea:

  1. Existing Server customers have access to a product that meets our needs at a reasonable price.
  2. New Server customers have a growth model so they can "start in the garage" and work their way up to a large enterprise install, without having to risk a disruptive migration from Cloud->DC along the way.
  3. Existing DC customers won't feel like Server customers are getting to buy the same functionality at a a lower price point than they had to pay.
  4. Atlassian gets additional revenue streams on a product they already support.

Win/win/win/win...

...unless what a lot of us suspect is true: Atlassian really has no future interest in customer-installed software at all, but are keeping their DC versions around to milk their larger cash cow enterprise customers for all they're worth?

Like # people like this

Well we should have seen the writing on the wall. Many tickets open for over 10+ years.

All we have seen in the cloud offering is non-stop UI changes and limited features. The UI has now gone full circle to the horizontal bar at the top again, except the handy "Issues" button is nowhere to be seen. In an ISSUE TRACKER, that's kind of an important feature to put in an obvious place.

You've lost a customer. The only thing to see now is how quickly I can get out.

To be 100% honest, I've given their cloud services so many looks over the last couple of years expecting that for sure, by now, they can basically match what their servers counterparts can do but I come back disappointed each time.

I've given their cloud services many chances but the problem with it in our case is just the simply inherent issues with putting particular data on the cloud and the lack of being able to customize things which are important to us. It's definitely useful in a lot of cases but it's just not for us. And after 2/2024, I guess that means Atlassian products won't be for us any longer. :(

Like # people like this

Saturday's announcement was not expected. We have been a customer since a few years and we picked Server from the get go precisely because it offer a straightforward "in-premises" solution.

You have now pushed us to start a migration process and, as you are only too aware, this means reassessing the solutions from your competitors.

We're not interested in your Cloud solution and the Data Center pricing is unreasonable.

Also, reading your e-mail announcement, what comes across is worrying. It suggests Atlassian cannot afford to continue to develop the Server product and puts in question your long-term sustainability.

Like # people like this

Hello to everyone at Atlassian and to everyone who has expressed their opinion concerning Atlassians' decision to change its server and Data Center offerings policies.

I am a Network and Systems Engineer in the IT Department of a major US college.  One of my responsibilities is the installation, configuration, and maintenance of several Atlassian products.  These products include three (3) instances of Atlassian Confluence Server 7.0.2 and one (1) instance of Atlassian Project Management Software 8.4.1.

All four of the above Atlassian products are fully licensed. Each of the Atlassian Confluence Server 7.0.2 instances are licensed for various numbers of users.  The Atlassian Project Management Software is licensed for an unlimited number of users.

We have successfully used these applications for close to 15 years now.

Each of these applications have been specifically configured to accommodate our internal network needs as well as the needs of our users, both public and private.  Our community has very specific configuration needs for each of the four Atlassian applications in particular the need to provide a secure environment.

Periodic updates of our Atlassian products are essential to maintaining functionality, addressing bugs, and ensuring a secure environment.  The loss of the ability to perform these upgrades locally and to allow us to locally configure our Atlassian applications would be catastrophic!  Any migration to cloud based instances (and the subsequent implementation of the stated server and Data Center changes) of our Atlassian applications would be unacceptable!

To add insult to injury it also looks like you are making drastic price increases for no reason other than to improve your bottom line and to bring the expectation that your cloud product pricing makes sense.

If Atlassian continues to insist that the "recently announced some changes to our server and Data Center offerings to sharpen our focus as a cloud-first company and deliver the world-class cloud experience you deserve" be put into effect would unfortunately cause our organization to altogether discontinue the use ALL Atlassian products and force our organization to seek alternative products to meet our needs!

Just making this announcement has reduced our trust in Atlassian greatly.



Bryan,

I hope it's clear to you that we don't all want your vanilla public cloud. We want to continue to run our private clouds! I work with a non-profit and we just spent a small fortune developing a custom application that runs on Confluence server and your decision hurts us tremendously.  As the others have said, this appears to be a decision that mostly benefits you and leaves your client's who have invested in private clouds in a very difficult position.  I feel what you are doing is morally wrong and even shameful. The public "cloud" will NEVER be a fit for many organizations who need far more flexibility that public cloud providers will be able to provide. I think that history will also show that the benefits of public clouds are massively overhyped. There is really nothing special about connecting to a server in your datacenter vs our datacenter and we lose an incredible amount of flexibility. I say this as someone with over 25 years of experience in IT and high tech. My suggestions is that you seriously reconsider this decision. 

Regards,

John Newcomb 

I agree with you.  I also think that if you read the Atlassian boiler plate responses that it's clear they anticipated our responses and simply do not care about our position.  They will continue to beat their drum on how life will be so much better for everyone moving into their infrastructure ignoring that we do not want it, never will.

Even if they reverse, its too late in my opinion.  I think they won't continue development with on-premise versions and will continue with thier wonderful "Cloud"  business model.  The price increase is just too much, it feels like punishment for going against their new movement.  I find it hard to believe they read our responses and come to a conclusion they got it wrong that we would just go along with moving into their cloud.  Instead, it's more likely they don't want us on-premise customers anymore and want the future cloud customers only.  Sorry Atlassian, I'm not giving you this control.  All within the scheme and mantra of lowering development costs, support and consolidating what is to become a huge mess.

As others have said, other new options may come from this which will probably be better for us in the long run.  I sure hope so.

I'll lastly say that this last year I've been de-clouding customers from the big three massively giving their freedom back of lower cost, security, performance and reliable uptime.  All which degrade over time with them and "control panel experts" diluting important system architechture skills.  But, thats for another conversation.

-G

Like # people like this

I have already sent some private messages and opened a support ticket but, due to the draconian nature of your cloud Acceptable Content Policy, and the fact that you retain the right to delete our content without notifying us, Atlassian as a company is preventing us from moving to cloud even if we wanted to (and we don't).

This change disrupts several years of our future growth plans and, unless things are done to address them and help cushion the losses from server investment we've paid, it feels like you are quite literally kicking us from your platform.

We would love to continue using Atlassian products, but it feels like you quite literally will not let us. You were one of the last bastions for professional self-hosted solutions, and you are giving that up.

Also, what happens if our company grows and we need to upgrade to the next level of license after the January date? Will we be forced to buy cloud?

 

Edit: I see, it ends in 2022 unless you own starter licenses.

Like Taylor Huston likes this
7 votes
JoeSu I'm New Here Oct 18, 2020

It is a disaster. I can hardly believe it. How can you do this to your customers? Unfortunately, money ruins everyone.

I speak for a company that's been using Atlassian products for years. While you obviously have put a lot of thought into this announcement (and the over all strategy), it is a huge disappointment, from a customers point of view, nonetheless. I'm afraid I have to agree with many others as they have voiced their concerns (to say the least), that this is very unfortunate.

If Atlassian actually follows through, we will migrate to different platforms from different vendors. After all, if Atlassian pulls the plug on local instances once, why trust them not to do it again in the future?

Suggest an answer

Log in or Sign up to answer
TAGS

Community Events

Connect with like-minded Atlassian users at free events near you!

Find an event

Connect with like-minded Atlassian users at free events near you!

Unfortunately there are no Community Events near you at the moment.

Host an event

You're one step closer to meeting fellow Atlassian users at your local event. Learn more about Community Events

Events near you