Geoff Shang wrote:> Kelvin Chu wrote: > >> After reading the docs, I was under the impression that icecast2 does >> not do on demand relaying? > > > Oh, on demand *relaying*! different kettle of fish. > > On demand relaying is available in Karl's branch and has just been > checked into the mainline development code. > > Geoff. >Tarballs of Karl's branch can be found at http://mediacast1.com/~karl/ Stephen LiveIce Project http://liveice.sf.net/
On Fri, 10 Jun 2005, ml wrote:> Tarballs of Karl's branch can be found at http://mediacast1.com/~karl/I wouldn't recommend kh8 or kh9 - I'm going down trying them until I hit kh3 (the last one I know works) or I find one that doesn't have problems with on-demand relaying. Most of the time it doesn't actually manage to do the relay (ending up with a timeout and/or a 404), and there are other oddities I'm finding it hard to pin down. kh9 has the added nasty bug over kh8 in that it doesn't seem to count relay listeners properly - so it never increases listeners above 0 to decrease them to 0 again and thus kill the on-demand relay, so relays started never stop. I've only just started looking at this in the past half hour, so unsure what exactly is happening. Now to try kh7 ... As said, I know kh3 definately works, as that is what I was using before I decided to look at later ones. Karl ?
On Fri, 10 Jun 2005, gARetH baBB wrote:> As said, I know kh3 definately works, as that is what I was using before I > decided to look at later ones.I ended up on 6a working for me (so far).
On Fri, 2005-06-10 at 10:11, gARetH baBB wrote:> On Fri, 10 Jun 2005, ml wrote: > > > Tarballs of Karl's branch can be found at http://mediacast1.com/~karl/ > > I wouldn't recommend kh8 or kh9 - I'm going down trying them until I hit > kh3 (the last one I know works) or I find one that doesn't have problems > with on-demand relaying.you know, you could actually tell me about any issues that crop up, having a report like this under the subject of icecast1 documentation is easily missed.> Most of the time it doesn't actually manage to do the relay (ending up > with a timeout and/or a 404), and there are other oddities I'm finding it > hard to pin down.on-demand relays can fail with those cases, if the source of the relay is not available. Again specifics will help, I've had little in the way of reports of problems.> kh9 has the added nasty bug over kh8 in that it doesn't seem to count > relay listeners properly - so it never increases listeners above 0 to > decrease them to 0 again and thus kill the on-demand relay, so relays > started never stop.not stopping is a new one, if you have a test case for this then fine. I had removed the pending queue recently, and it looks like the count was off by 1. Fixed in svn now, I don't see the count not increasing and decreasing at all though, what is the case you're using.> I've only just started looking at this in the past half hour, so unsure > what exactly is happening. > > Now to try kh7 ...again, do let me know. might save you a lot of work. karl.