Has SageTV 9 Gone Linux Only

[emotionnotion]

Guys,

can't help noticing a lot of noise around Linux and not much in the Windows domain with the new version of SageTV.

Are we leaving Windows behind as if we are I have some serious thinking to do given my core system is Windows Home Server based.

Would appreciate an idea of where this is REALLY going!

[Fuzzy]

Not much discussion of windows, because not much has changed there.. more people are interested in the linux build at this point because it can run on a 64-bit JVM, so for those with a lot of clients, it's a godsend. I know my next server rebuild will probably be linux based because of it.

[emotionnotion]

Thanks Fuzzy.

I have 3 HD300s and 2 HD200s and so, this may make sense for me. Even if some of them start to die, I can create mini-PC extenders fairly easily.

If we could make the software and update experience on the server side as easy as the HD300 software and updates, we will be most of the way there.

[stuckless]

Originally Posted by emotionnotion

Thanks Fuzzy.

I have 3 HD300s and 2 HD200s and so, this may make sense for me. Even if some of them start to die, I can create mini-PC extenders fairly easily.

If we could make the software and update experience on the server side as easy as the HD300 software and updates, we will be most of the way there.

There is also the Android MiniClient, which isn't perfect, but, runs pretty well for most cases. I've replaced 2 of HD300 extenders with NVidia Shield Android TV units.

As for windows, someone needs to own it, and not sure we have people even trying to make the windows code 64bit. Jusjoken has created the windows installer which will install V9 32bit... and I think the installer is for Server, Client and Placeshifter.

Most of sagetv "features" is java based, so, any changes there will work on 32bit or 64bit (linux or windows). At this point most of the "native" code changes for linux have been to fix issues with moving to 64bit.

The linux build system is pretty easy to run, and can run unattended, so I think the reason for more linux builds, is just that it's easier. The windows build system seems to have all kinds of dependencies on VS tooling and it's not like that stuff can be installed easily.

At some point, I fully expect that someone will just "own" the windows native parts and rebuilding, etc, and push out new native builds when things change. Community just isn't that large yet.

[KarylFStein]

I'd send a donation to someone who gets 64 bit Windows working . Ever since moving to Gemstone I get JVM heap space errors every once in a while. If I were building a new Sage server I'd go Linux, and will probably move that way in the future.

Getting the JAR file to build on Windows was pretty easy with the instructions in the forums. This allowed me to get a fix installed on my system even though a new Windows installer version wasn't made. If there was a base 64 bit Windows install available it may not have to be updated much at all. People can install that then figure out other ways to handle JAR updates. I think most changes take place to the JAR.

[Fuzzy]

Originally Posted by KarylFStein

I'd send a donation to someone who gets 64 bit Windows working . Ever since moving to Gemstone I get JVM heap space errors every once in a while. If I were building a new Sage server I'd go Linux, and will probably move that way in the future.

Getting the JAR file to build on Windows was pretty easy with the instructions in the forums. This allowed me to get a fix installed on my system even though a new Windows installer version wasn't made. If there was a base 64 bit Windows install available it may not have to be updated much at all. People can install that then figure out other ways to handle JAR updates. I think most changes take place to the JAR.

There simply aren't any really significant window programmers in the community at this time to pursue any sort of bounty. We've got a few who can get around Linux well enough, and plenty of Java experienced guys. But it's the native windows libraries that are the limiting factor in moving the windows server to 64-bit, and there may be complication with any native tuner interaction as well, since not all tuner interfaces are 64-bit either. For this reason, the 'quickest' way to get a 64-bit server running in windows would be to strip it down to be just a bare JVM, that doesn't really have any of the native capabilities, then all tuning would be done via network encoders... but once you do that, it ends up being just as easy to run a Linux VM on your windows host and run the 64-bit JVM there...

[brandypuff]

I'm running Windows 7 x64 with the 32-bit V6 SagetV and plan to upgrade to v9 one of these days when i have some time. I don't ever expect to move to Linux as Sage is running on a centralized media and backup server and don't want to experience the pain of migrating to Linux after running solid for almost 10 years on Windows

[Fuzzy]

Originally Posted by brandypuff

I'm running Windows 7 x64 with the 32-bit V6 SagetV and plan to upgrade to v9 one of these days when i have some time. I don't ever expect to move to Linux as Sage is running on a centralized media and backup server and don't want to experience the pain of migrating to Linux after running solid for almost 10 years on Windows

And you are not 'losing' sage by staying windows only, so you shouldn't worry about it. There hasn't been any non-windows development with the exception of the linux 64-bit build, and there is no intention of moving away from windows. If 1gb heap has been enough for you in the past, there's no reason to move to Linux.

[valnar]

I also have a single Windows 7 "server" at home that runs a host of other applications I wouldn't want to find on Linux. I use a HDHomeRun, HD-PVR, USB-UIRT and other things for tuning that just works in Windows. For now I see staying on v7 for the foreseeable future.

After the dust settles on the the new Sage v9, maybe somebody could make a Linux ISO image that includes all the various add-ons most people use? 'Like Mythbuntu is to MythTV? That would be the only way I could ever convert.

[Fuzzy]

Originally Posted by valnar

I also have a single Windows 7 "server" at home that runs a host of other applications I wouldn't want to find on Linux. I use a HDHomeRun, HD-PVR, USB-UIRT and other things for tuning that just works in Windows. For now I see staying on v7 for the foreseeable future.

After the dust settles on the the new Sage v9, maybe somebody could make a Linux ISO image that includes all the various add-ons most people use? 'Like Mythbuntu is to MythTV? That would be the only way I could ever convert.

I'm confused.. Windows vs. Linux should have absolutely no bearing on v7 vs. v9.

[stuckless]

Originally Posted by valnar

I also have a single Windows 7 "server" at home that runs a host of other applications I wouldn't want to find on Linux. I use a HDHomeRun, HD-PVR, USB-UIRT and other things for tuning that just works in Windows. For now I see staying on v7 for the foreseeable future.

After the dust settles on the the new Sage v9, maybe somebody could make a Linux ISO image that includes all the various add-ons most people use? 'Like Mythbuntu is to MythTV? That would be the only way I could ever convert.

I highly doubt that we'll see an ISO or OS release dedicated to SageTV. I doesn't make sense... especially given that Docker is much more convenient way to accomplish the same thing.

I would hate to see SageTV go Linux only (even though I use only linux, myself)... but until Windows people start taking owership over the parts, then it might very well someday be Linux only... since I have no interest in spending my time on the windows parts.

Meanwhile, there are sagetv 9 installers for windows... so you can upgrade to Sage9... The challenge will be that you'll likely never see a 64bit windows version.

And I can certainly feel you pain points with Linux... I feel the same way about windows.

We need a Google Summer of Code to get some young kids involved in windows build system. Someone should look into that.

[cat6man]

Originally Posted by valnar

I also have a single Windows 7 "server" at home that runs a host of other applications I wouldn't want to find on Linux. I use a HDHomeRun, HD-PVR, USB-UIRT and other things for tuning that just works in Windows. For now I see staying on v7 for the foreseeable future.

After the dust settles on the the new Sage v9, maybe somebody could make a Linux ISO image that includes all the various add-ons most people use? 'Like Mythbuntu is to MythTV? That would be the only way I could ever convert.

i understand where you are coming from.
my experience is that i've left windows behind and moved to linux and have only found 1 bottleneck with a windows program (see below). i found i could run all of the windows programs i needed using either 'wine' alone or with 'play on linux' to install the windows *.exe files.
for esample, critical programs for me include dBpoweramp, tagsscanner, RootsMagic, Textpad, etc, all of which are windows only and run great on my linux machine............in addition to the sagetv linux server, i've moved my main home pc to linux rather than go to windows10............never ever going back

you also might consider running a virtual machine so that linux and windows can be running simultaneously, linux for sage and windows for your normal windows server.........reports are that sage runs great on a virtual machine.

note: the HDtracks downloader seems not to register JRtools.exe so i can't buy/download music from their site.............anyone know how to get that running under linux?

[SageWizdom]

Originally Posted by valnar

After the dust settles on the the new Sage v9, maybe somebody could make a Linux ISO image that includes all the various add-ons most people use? 'Like Mythbuntu is to MythTV? That would be the only way I could ever convert.

I can see doing this. While a docker image works for many folks, I like the idea previously mentioned of having a Raspberry Pi Sage server. In that case (or similar) there would be a standard image that loads, validates hardware, runs through the config and "just works." I would not expect that just yet though. If you want to try Linux, there are a couple install walkthroughs and install scripts... but it is time to update them again (oops, I need to get on that).

I'm definitely going to look at the list above of windows tools...

[wayner]

I love Raspberry Pis and I own a bunch of them, but I am not keen of the idea of a Raspberry Pi server. I don't think it can handle the I/O needed for something like Sage where you can have multiple video streams going to/from hard drives - don't forget that it doesn't have a SATA interface. That is especially true if we are going to ever be able to move to 4K which would increase the bitrate of files and I/O demands. And I don't know that Pis would work for transcoding which may be required in some instances.

[boonest]

What about a Raspberry Pi extender to replace a HD300?

[Fuzzy]

Originally Posted by boonest

What about a Raspberry Pi extender to replace a HD300?

That does make more sense for a Pi's hardware than a server does. It would be a good target for it, but when you compare it to, say, and android tv type device that can also do netflix AND sagetv, it seems to lose out.

[SageWizdom]

I would agree. The ability to support online video content without having to write another ton of apps is a plus. I know it was previously mentioned the RPI as a server. I would be curious to see where it gets stuck, or others thoughts on a low power server.

The only numbers I can seem to find are 20-25 MB/sec for the USB or SDCard. However, I can only find this info on forums and not on any official site. I have not dug into chip specs which is where I imagine it comes from.

Any thoughts on alternative low power / small options. For example, if I want to stick one of these in my parents/friends/relatives house and only check in remotely when they are having problems... thoughts on a good small box to stick in a closet somewhere with some sort of network recorder?

[sdsean]

Well I'm pretty dedicated to windows and I'm not really wanting to setup new hardware or a dual boot for linux. . .

But I'm desperately wanting to have a 64bit JVM. . . I have 3 extenders (HD300s) and I cannot get them to all run together with the few plugins I do have and a 32bit JVM.

I do like the idea of using a VM though. . . but would I be able to get tuning and what not to work right? I'm using serial port turning to DirecTV STBs, and have 4 HD-PVRs. . .

If I could make that work. . .that would be awesome. . .

But until that happens. . .its windows. . .

Fuzzy. . .

But it's the native windows libraries that are the limiting factor in moving the windows server to 64-bit, and there may be complication with any native tuner interaction as well, since not all tuner interfaces are 64-bit either.

What native windows libraries specifically are you referring too? In the SageTV code? or windows libraries themselves (DirectX? MFC? etc)?

[sdsean]

I see the win32 shell hook code in git. . but that should be pretty simple to port. . . there's some stuff throughout the code using the win32 version of getting the current date/time but that's really easy to fix. . .

[Fuzzy]

There is quite a lot of native code in sagetv that can't currently be called from a 64-bit JVM. MOST the native code in there is for the client, so a lot could be stripped out to make a 64-bit server only compile - but there are some server related parts that are native - the Power Management interfacing, checking free disk space, reading and writing windows registry keys, the BDA and DirectShow capture system, channel changing (USB-UIRT, for instance) - these are things that most people would require their server to be able to do, and they are 32-bit native libraries (called via JNI). Some of this can be worked around architecturally by using only network encoders (honestly, this would be my preferred architecture for any sage work - but I like total modularity) - others might be able to be easily ported to 64-bit. Some might be able to use external utilities instead of linked libraries. The real problem is we don't really have anyone active in the community comfortable enough with windows development in general, and JNI specifically, to really even start to sort it out - I'm certainly not the guy to go to, I've just spent a lot of time reading through the code to try to get a deeper understanding of the challenges.