Network Encoder HowTo

[bcjenkins]

Greetings,

I read through the manual:

SageTV can be used to control recordings made with an encoding device that is installed on
another computer on the same network. Two full installations of SageTV are used for this
purpose:

For the installation of SageTV that is going to be the network encoder, set this property:
enable_encoding_server=true. Start this installation of SageTV. Note: To also use
SageTV on this system as a client, start a second instance of SageTV.exe, using the ‘-
client’ command line parameter. See Appendix D: SageTV Command Line Parameters.
For the main installation of SageTV (the one that is to be the server for all your clients),
set this property: network_encoder_discovery=true.Start SageTV.

Note: Each installation of SageTV will require a separate registration license.

Note: The SageTV server will need access through any software firewall for port 8271
(UDP), while the network encoders will need access for port 6969 (TCP). These ports
are the same no matter how many network encoder devices are in use, on one or multiple
computers. See Firewall Settings for more port information.

However, it is not working. I have a couple of questions I am hoping I can get answers to.

1. On the encoding server, do you set up any adapter, guide, or channel mappings at all?
2. If so, do you set up the master in the same fasion, or does the master simply pass along a channel request?

Thanks!

B

[Opus4]

Some of your recent posts were in the Linux section of the forum... by any chance, are you trying to use a Linux system as one or both parts of the the network encoder setup?

- Andy

[bcjenkins]

by chance I am, as the slave

The setup docs aren't clear as far as platform requirements, and the Linux FAQ says it can be done.

B

PS, I am asking from the master server point of view which is why this is not posted in LINUX forum.

[Opus4]

There is a known issue when using a Linux setup as the network encoder for a Windows main server. Right now, it has to have Linux on both sides of the network encoder setup w/the smae file path that they can each see directly.

- Andy

[bcjenkins]

For my own edification, it sounds like the guide on both sides needs to be configured the same as well as recording paths? I couldn't tell from the other threads how specifically to set up the master server. Isn't it the one responsible for writing files to the file share?

Supposing both sides were the same. Either windows or linux, etc.

TIA

B

[stanger89]

If I could guess, I'd say the server just passes the recording directory on to the slave, if these aren't the same OS, paths are different format (ie //server/share doesn't work on Linux, at least not out of the box).

I think there would need to be some mapping done on one side or the other to map between linux paths and Windows paths, and that's not being done at the moment.

[GKusnick]

All video sources, channel lineups, and recording directories must be configured on the primary server, since that's where the unified EPG lives and where recordings are scheduled. Whether they also need to be configured in parallel on the network encoder I don't know, having never done it, but at a guess I don't see why they would. In theory all the encoder should need to do is tell the primary what capture devices it has available (per Device Manager), so they can show up as configuration candidates on the primary's Video Sources screen. After that, the primary server manages everything.

[bcjenkins]

Originally Posted by stanger89

If I could guess, I'd say the server just passes the recording directory on to the slave, if these aren't the same OS, paths are different format (ie //server/share doesn't work on Linux, at least not out of the box).

I think there would need to be some mapping done on one side or the other to map between linux paths and Windows paths, and that's not being done at the moment.

Can we use CIFS syntax for file store locations? I'll try this in the morning when noone is trying to watch TV. http://www.microsoft.com/mind/1196/cifs.asp. If this is indeed the obstacle, I could also come up with some other ways of mapping the directories through the use of junction links too. Assuming I could place one directory per path and link it to the store with junction links or smb mount points.

Originally Posted by GKusnick

All video sources, channel lineups, and recording directories must be configured on the primary server, since that's where the unified EPG lives and where recordings are scheduled. Whether they also need to be configured in parallel on the network encoder I don't know, having never done it, but at a guess I don't see why they would. In theory all the encoder should need to do is tell the primary what capture devices it has available (per Device Manager), so they can show up as configuration candidates on the primary's Video Sources screen. After that, the primary server manages everything.

So the only thing which should be done on the slave is set the property for running slave and make sure the adapter can be tuned? Since I already have my guide mappings done on the slave, I can copy them down to be redone on the master. I am using linux for QAM just to test.

B

PS. Thanks for the replies

[stanger89]

Originally Posted by bcjenkins

Can we use CIFS syntax for file store locations? I'll try this in the morning when noone is trying to watch TV. http://www.microsoft.com/mind/1196/cifs.asp. If this is indeed the obstacle, I could also come up with some other ways of mapping the directories through the use of junction links too. Assuming I could place one directory per path and link it to the store with junction links or smb mount points.

You can do CIFS in linux via Samba, but last time I tried the syntax was different for paths, and I think that's the inherent problem currently.

[Opus4]

Originally Posted by bcjenkins

For my own edification, it sounds like the guide on both sides needs to be configured the same as well as recording paths? I couldn't tell from the other threads how specifically to set up the master server. Isn't it the one responsible for writing files to the file share?

I'll go along with what was posted above... you should just be able to set it up with those few properties changes that the manual mentions, then configure the source via any UI connected to the main server. (There is no UI connected to the network encoder, so you wouldn't be configuring the source there anyway.) I was never told that there was a need to use network paths with 2 Windows systems, so it wasn't listed as a requirement in the manual -- the network encoder would send it back to the server, which would save it (or something along those lines).

There is one caveat here: I haven't set up network tuners either, so I'm just passing along the instructions I was given, as I understand them. The manual section for this was proof-read by someone who knows very well how this all works, though.

- Andy

[bcjenkins]

If SageTV were to use http://jcifs.samba.org/ as a loadable module, theoretically it could frame the request the same on both platforms? I am assuming this would require something more than loading it as a module, but maybe someone with better working knowledge of file i/o in SageTV could chime in.

B

[bcjenkins]

mkdir /opt/sagetv/server/\\\\sagetv-server\\sagetv

Creates a directory \\sagetv-server\sagetv in the root of the sagetv server directory on the linux box. This matches the name of my unc share. I then mounted:

mount -t cifs //sagetv-server/sagetv /opt/sagetv/server/\\\\sagetv-server\\sagetv -o user=username,pass=password,domain=domainname

And can now traverse the server and created a directory. Note the use of CIFS fs instead of smbfs in the mount command. Apparently smbfs is now deprecated.

Based on the conversation above, I need to reconfigure my channel lineups etc on the Windows box and strip down the config on the linux box. right?

Back to fiddling..

[stanger89]

Remember, the conversation above is all speculation.

[bcjenkins]

Right, so is my ability to make this work.

[bcjenkins]

I think I figured it out mostly..

I will have to test later, unless someone else can confirm in the meantime. This would not be an appropriate long term solution as it, but it might get folks past the hurdle.

Windows would allow $FOO as a directory name. In linux $FOO would be treated as a variable. So if FOO was exported to /var/media/tv/ (trailing slash important) and /var/media/tv was mounted to your windows share for tv recordings.. It should work. The even better part is that $FOO\recordingfilename.mpg from Windows is what linux needs to make this work. I created a dir in linux called tv and did:

export DIR="tv/"
echo hi > $DIR\hello.txt

which resulted in a file called hello.txt containing hi in the tv directory.

The only items I was unable to confirm this morning:

1. Will SageTV allow the use of just a directory name for the recording directory?
2. Where will the directory need to be placed? I am assuming it goes where sagetv.exe resides.
3. Can Superman stop the speeding train in time?

Thoughts? My test later will be with a server setup on my son's pc. If it works, I will use junction link magic to point the directories to their actual locations. btw - If you are running Vista and presumably server 2008 junctions are built in to the OS now.

B

[bcjenkins]

Thought about it some more..

Here's what I can do

On windows set path to be:

\\server\share\a$VAR
\\server\share\b$VAR

etc.

On linux create directory \\server\share\a... export $VAR = "/" mount \\server\share\a\$VAR to new dir.

Benefits: allows me to keep unc paths for windows clients.

Now the question is: Where will SageTV on linux think the dir \\server\share\a is located?

B