Jetty 9 (to Upgrade or Not)

[stuckless]

I've spent way too many hours looking at different options around upgrading Jetty.

First off.. As Java advances, some versions of Jetty will ONLY work with Java 8+... And the current version Jetty 6 has known issues with some java apis when running on Java 8. So if we don't upgrade jetty... we may be stuff on Java 7 for a long time (which I don't like).

One problem is that no matter how we upgrade, we'll never be able to do it without breaking all the existing webapps. Some of them like the ./apps will need to rewritten, and others like nielm_sagetwebapp will need to be tweaked.

I been trying to find a way to run the "old" and "new" in the same setup, but that's never going happen. I've even tried to running a jetty server (old) and using tomcat (new) instances and that almost works, but I see it causing many more issues.

One of the reasons why all the existing webapps need to be updated is that Jetty, at some point, changed owners, and the packages were all renamed... so the Context xml files that we use for Jetty 6 will not work in Jetty 9.

So, the upgrade strategy would basically be a new Jetty 9 instance with a completely different set of configuration files, etc, so "upgrading" to jetty 9 would break existing web apps (until they are updated) and will require a reconfiguration of the server configuration... ie, ssll, user/pass, etc.

Obviously this does not affect SageTV 7, but, would affect people running the Open Source version... My feeling is that if you are on the Open Source version (which is pretty stable in my mind), you'll likely be more tolerant of breakage... since, you are on the bleeding edge

[Bagal]

I say go for it, better to get the pain of upgrading out of the way in one hit than having to worry about different versions etc. going forward and as you say those on the Open Source version should be more tolerant of breakage anyway.

Now I just need to find some time to install and play with v9...

[KarylFStein]

I don't suppose anyone's willing to take over Sage Alert? Can you tell if it's just a "tweak" or a more involved process to get it on a newer Jetty version? Or maybe add email as an option to the Pushbullet plugin? I mostly care about getting system messages via email.

[Steve20A]

Do it once. Do it right. Consolidate the pain - as you say, early adopters are a hearty lot - before v9 becomes the predominant version in use.

I moved from v7 to v9 after a lot of testing and haven't updated since. Maybe I'm what you would call a timid early adopter, but still, I say go for it.

[stuckless]

Originally Posted by KarylFStein

I don't suppose anyone's willing to take over Sage Alert? Can you tell if it's just a "tweak" or a more involved process to get it on a newer Jetty version? Or maybe add email as an option to the Pushbullet plugin? I mostly care about getting system messages via email.

I think adding email to the pushbullet plugin would be somewhat easy... but also, I think the Sage Alert (or any of these web apps) would just require a recompile (maybe) and an updated context.

Personally, I think I'd like to incorporate some of these features into the "core" web ui... ie, as part of the configuration allow you to register for alerts, etc, and move those plugins as being something that is core configuration.

[RBraverman]

Looks like the discussion fizzled quickly for some reason.

If you are going to do this, it would require (in my VERY humble opinion) people who have test servers/clients.

Even though I'm on 9 and have beta'd more software and hardware than I care to remember - this type of app doesn't lend itself to "playing/testing".

And considering there are many who are updating and can not get the Jetty Server to work ....

All that being said, I agree with the prevailing opinion - if you are going to take this on, do it once and do it right - an old saying I remember when I was a line mechanic for Porsche, "You never have the time or money to do right the first time, but you ALWAYS seem to have both to do it over."

If I can be of service, I've a test bed machine that I can easy set up, so I wouldn't cry very loudly and through a tantrum if everything

[stuckless]

Originally Posted by RBraverman

Looks like the discussion fizzled quickly for some reason.

If you are going to do this, it would require (in my VERY humble opinion) people who have test servers/clients.

Even though I'm on 9 and have beta'd more software and hardware than I care to remember - this type of app doesn't lend itself to "playing/testing".

And considering there are many who are updating and can not get the Jetty Server to work ....

All that being said, I agree with the prevailing opinion - if you are going to take this on, do it once and do it right - an old saying I remember when I was a line mechanic for Porsche, "You never have the time or money to do right the first time, but you ALWAYS seem to have both to do it over."

If I can be of service, I've a test bed machine that I can easy set up, so I wouldn't cry very loudly and through a tantrum if everything

It's still being played with It's just that I have Phoenix, Android MiniClient, BMT (rewrite), and as such, with only a couple hours a week to spend on things... it's moves at a snails pace

[RBraverman]

Originally Posted by stuckless

It's still being played with It's just that I have Phoenix, Android MiniClient, BMT (rewrite), and as such, with only a couple hours a week to spend on things... it's moves at a snails pace

And to think, I thought all you did was SPEND EVERY WAKING HOUR fixing STV.

What's the re-write of BMT for? Maybe the "Service_Unavailable" huh, huh, huh?

Also a thought, if/when you ever get to tackling Jetty, maybe a list on a sticky of the all the major addons, have people vote on the ones that mean the most to them, then take oh, the top 3-5 and make sure they work with the re-written Jetty, before a release?

OR so that one could list the items that would be impacted.

As ALWAYS thank you for MASSIVE contributions!!!!!