On 9/26/07, Jeremiah Benham <address@hidden
On Tue, 2007-09-25 at 13:36 -0700, Christian Einfeldt wrote:
> I wanted to thank everyone who helped on this thread. We have cobbled
> together a solution that allows us to play GNU Denemo in the GNU Linux
> lab over edubuntu. Essentially, we require the students to start play
> sequentially, rather than simultaneously. It takes a wee bit longer,
> but it works.
Why did you not cc the denemo developer mailing list about this issue?
Sorry about that. I am a volunteer for a public middle school, and things are a little crazy here. Also, I went to the denemo IRC as a recall, and there was little activity. Or perhaps it was a forum, and there was little activity. At any rate, I didn't find it, and I am glad that you found me. Denemo is really awesome, and it has done incredible things for our public middle school. We are really grateful for Denemo.
I am a developer for Denemo.
Wonderful. How great to hear from you! Again, at the risk of sounding overly enthusiastic, you have done great things for us. Here is our school. It is a public middle school in inner city San Francisco. 75% of our students live in households below the federal poverty guidelines. Of the total 100%, 65% of our students are African Amercian; 17% are Latino; 10% are Asian; and 8% are Caucasian. So when you hear people talking about beating the digital divide, please understand that you are doing great things to help us beat the digital divide !
I rarely read Linux-audio-users. I usually
just scan over it now and then.
Yeah, so I am really glad that you happened to see my post!
What version of denemo are you using?
It is the version that comes with our edubuntu install. I am guessing it it edubuntu Feisty.
How have you temporarily solved the problem?
We found that the choking happens only when all of the students play their Denemo compositions simultaneously. If we take the students one-by-one, we don't have that problem. This does not mean that we have to wait for the students to finish before another student starts playback; rather, it means that we have student 1 start playback and then nod at the teacher when they hear sound; then the next student starts playback; and then the next student; and so forth. That is not as good as everyone being able to play their compositions at the same time, but it works for now.
Here is our hardware: it's a Pentium D 915 (dual 2.8 GHz) with 2 GB of RAM. We are using Cat 5 cable and two GB switches. About 18 of the clients are HP Vectra VL 420 towers or pizza boxes with 256 MB of RAM and about P4
1.2 Ghz chips. The others are Dell OptiPlex GX110 pizza boxes with 256 MB of RAM and about 800 Mhz chips
We sometimes have as much as 25 or 26 on the system at the same time. The physical layout of the network is:
Server > switch 1 (18 clients) > switch 2 (10 clients). We often will have only 23 clients running at the same time. The boxes are sitting on 5 rows of tables with 6 clients per row, for a total of 30 clients. The last 12 of the clients are behind switch 2. The first 18 clients are in front of switch 2. The first 18 clients are drinking from switch 1. Switches 1 and 2 are both GB switches. We believe that all of the clients have decent client NICs. We don't think the NICs are the problem.
I notice that when the server goes down, its shutdown menu says something about timidity not being configured properly. The actual message is something like "timidity is not configured yet. Enable ALSA squencer at /etc/default/timidity." I am actually just now getting the time to pursue that error message. The guy who built this system mentioned something to the effect that sound was a known problem over edubuntu thin client networks, but I have not yet had a chance to ask him about that further, as to what he meant. We were rather rushed in putting this lab up, due to chronic funding shortages in California schools. I will google that message now and get back to you.
Thanks for your interest, Jeremiah !