Version 6 migration path
Our main website is running V5.0 and I've resisted upgrading something that is working.
I can see a couple of features in V6.0 that would now justify upgrading, but am balking at the idea of stepping though multiple versions to get there.
Will V6 offer a simple upgrade path from V5.x or are we going to be forced through this insane multiple version upgrade process ?
I am sure you can give it a go. I have done a upgrade difference of 3 versions once and it was fine. But I would do stepped upgrades myself and I wouldn't expect, (or want the Sitefinity team to put time into), to support such a upgrade scenario.
Me, I upgrade all the time. I upgrade my customers for free because upgrading from version a to b is always fully supported and generally issueless. (Plus I get heaps of bug fixes and improvements and that only makes the customers happy)
I am not of the "if its not broke don't fix it." I have vb6 projects that ain't broke but who the smurf wants to be dealing with those now?
I like it to upgrading VS 2003 projects to 2010, (Of which I have done 30 in the last 4 months) Too much has changed and I upgrade version to version and have no problems. (If I try the big jump I do)
(And I think this applies to most apps and there upgrades. (We'll at least Microsoft based ones.))
Upgrading a SF project would be the same. You upgrade locally a few versions at a time and do one prod upgrade at the end.
I totally understand that in your environment, this may not be easy. But, if that is the case, I would encourage you to look at why that is and change things so going forward you can.
Not the answer you were looking for I know, but perhaps a change of thinking might be better here.
Dear MB
There is a saying never touch a running system. However somewhere in the future you will come to a point where the 5.0 will not work anymore.
I usually do upgrades for my clients once a year. Tomorrow I will upgrade an 4.4. site to 5.4 and yes with all the steps.
5.1 introduced the turning of of modules if I remember correct which saves me some RAM on the server. Which is great. 5.4 fixed some library issues (sorting) these alone are two things for me making a upgrade worth.
And if I ware you I would not expect any upgrades from 5.0 to 6.0 directly. Telerik has had the policy of supporting two major versions back in upgrading.
It's always nice to be up to date so good luck.
Markus
If they simply provided export and of course import of the module data, generic, news, events, lists, and module builder modules they would give a person the tools to simply start version jump upgrades going forward by creating a new project, compile in anything custom and then import the data. This would get you 90 percent of the way in less than an hour.
As for upgrade for free that is very generous considering you have to backup the database and original solution, get that backup live so you can upgrade it instead of the live site. Upgrade it, then make sure to upgrade all the files and configurations the Sitefinity project manager does not upgrade. Then you have to test everything, try each module, each widget to see if it blows up in your face or not. Hours of work because it is always like slow motion dealing with such heavy .net applications (Thank God for Lightning). This process usually ends up with some bug or missed configuration or file change that causes something to break with hours of troubleshooting to correct the issue if it can be corrected if it is a bug you have to figure out what to tell the customer.
Take that with a huge upgrade jump say 4.4 to 5.4, you can't do that all at once you have to step through not all but many of the versions in between. Same scenario for each upgrade step and I mean all the testing and correcting issues.
Moving forward the Sitefinity project manager needs to pull the data from the current site and inject it into a new fresh current version site. There is still all of the unknown custom work that Telerik can not be expected to anticipate but this will nail the core 90 percent of the upgrade. They would have to give us page/theme export import as well to really make this an option but even with out that I would extremely happy with module import and export for these upgrades.
When I think of companies managing a few dozen of these sites or more I wonder how they pull it off and stay profitable.
Time for some Walking Dead?
Dear Billl
I just took a 4.4 site to 5.4 this weekend. The update processes alon did take me 6 hours. Backing up, transfering data, cecking each version if we have problems or not. Did not run into any - so I was lucky to get away like this. No custom stuff though.
If you run your own site its not so bad. But as you mentioned if you got more then 10 sites this would amount easyely up to 1-2 weeks work.
However I do understand that exporting importing is very very complicated to not forget any of the chnanges you have made in structure and stuff.
You know 90% nailing is only doing 10% of the work. To clean up the 10% you probably need 90% of the time.
So for the time beeing as much as I love to be able to take 4.4 to 5.5 in one step - I don't quite see this possible and as always would have my focus (but thats just me) on fixing and polishing current versions over and over.
Markus
I upgrade 8 sites over a weekend and most of that was waiting for data transferees, drinking a lot of coffee and watching a lot of missed TV.
Why can I do this? Because I have a mind set to develop for maintenance. My look on things is that there is only one thing that is guaranteed with software and that is change. So I ensure that the way I develop that I can change with out worry. (Well, without a lot of worry).
We should look at why our upgrades take so long and change it.
For example I automate all my backups and before I do a upgrade I tell the client not to make any changes during a certain period. Thus I can use the automated backups and don't have to kick off and wait for a manual one.
My upgrade and deployment plans are the same and methodical. Yes, I had to spend time experimenting and working it all out but it is akin to Unit Testing. Its more work up front but its great having the security down the track.
Don't get me wrong. Customers sometimes force us to do stupid things even when we label it as a stupid thing to do. So getting this utopian environment isn't always achievable. But our jobs are generally to simplifier and automate other peoples businesses yet we often don't do it for ourselves.
And because I am in this state I of course want Telerik to focus effort on current bugs and polish along with new and refreshed features.
Bill and All,
There may be some light at the end of the "new project + data imported from old project" tunnel.
In the thread below, Victor Velev, from Telerik, posted a link to a PITS that we can vote on:
I'm going to upgrade to 5.4 w/in 2 weeks. That said, it'd be great if Telerik can make this work for version 6.
Gregory
Direct PITS link
http://www.telerik.com/support/pits.aspx#/public/sitefinity/14439
Vote on it if you haven't already!
I voted and will hit PITS tonight to vote on a few other ones