Hey Mark,<br><br><div class="gmail_quote">On Tue, Mar 29, 2011 at 11:01 AM, Mark Lord <span dir="ltr"><<a href="mailto:mythtv@rtr.ca">mythtv@rtr.ca</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex;">
<div class="im">On 11-03-29 10:53 AM, Mario Limonciello wrote:<br>
> On Tue, Mar 29, 2011 at 7:05 AM, Ryan Patterson <<a href="mailto:ryan.goat@gmail.com">ryan.goat@gmail.com</a><br>
</div><div class="im">> <mailto:<a href="mailto:ryan.goat@gmail.com">ryan.goat@gmail.com</a>>> wrote:<br>
><br>
> On Mon, Mar 28, 2011 at 1:31 PM, Mark Lord <<a href="mailto:mythtv@rtr.ca">mythtv@rtr.ca</a><br>
</div><div class="im">> <mailto:<a href="mailto:mythtv@rtr.ca">mythtv@rtr.ca</a>>> wrote:<br>
> > Q: Do you (re)build mythtv from source code? If so, why?<br>
><br>
> I build from source because I hate the "enhancements" the ubuntu<br>
> package maintainers implement.<br>
><br>
> Do you have some specifics? We're quite open to making changes if there are<br>
> things people don't like.<br>
<br>
</div>I won't speak for Ryan, but all of those "*.real" wrapper scripts<br>
simply drive me bananas.<br><br></blockquote><div><br></div><div>Is there a specific problem you're hitting with them? It could easily be a bug that just needs to be fixed with them. </div><div><br></div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex;">
I know why they are there, for the theoretical "non-technical users",<br>
and I don't expect them to disappear, but I don't want them here! :)<br>
<br></blockquote><div><br></div><div><meta http-equiv="content-type" content="text/html; charset=utf-8">Ideally at some point I would like to see them go, but for now they do manage a few permissions related problems that came up in the past as well as some nicities such as respawning the frontend if it exits non-zero (eg crashes).</div>
<div> </div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex;">
Speaking of Mythbuntu -- it would be Really Nice if somebody there could<br>
take a look at the Upstart startup/shutdown sequence. It mostly works<br>
by accidents of timing, and doesn't work at all in the system I have here,<br>
with the rootfs on a very very fast SSD. That throws the timing way off.<br>
<br>
Examples:<br>
<br>
(1) The box here boots/runs from SSD, but has a 2TB mechanical drive<br>
for recordings and videos. Quite frequently, about every third to fifth boot,<br>
something buggy in the startup doesn't "see" the mechanical drive, and starts<br>
up the system without mounting it ("auto" in /etc/fstab). I've now removed<br>
the "auto" from /etc/fstab, and added my own script to mount it later on.<br>
I also had to add some delays in the initramfs so that udev wouldn't miss<br>
seeing that device. Ugh.<br></blockquote><div><br></div><div>Have you filed a bug on upstart yet about this? It sounds like a core problem that some of the upstart developers would need to provide some input on.</div><div>
</div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex;">
<br>
(2) mythbackend frequently start before networking is up, so it doesn't find<br>
the hdhomerun network-based tuners. Again, I've bypassed this by writing my<br>
own scripts to start mythbackend now, and changed the sequence to require<br>
networking.<br></blockquote><div><br></div><div>This actually can be fixed now for natty finally because there is a Network Manager upstart job. We just need to make sure not to break maverick and lucid in the process. The same package gets built on all three of them.</div>
<div><br></div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex;">
<br>
(3) At shutdown, mysql exits before mythbackend is shut down. Duh..<br></blockquote><div><br></div><div>This is an unfortunate side effect that upstart doesn't support optional requirements for starting jobs. What we need is something that is "if this job exists, it should block startup of the next job until it starts, but if it doesn't, then don't". I requested this from the main upstart developer a while back, but it <i>still</i> hasn't landed. Creative things like just having a pre-start script that pings mysqld until it's running might end up have to be the solution that's used. Was your workaround for this any more creative? </div>
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex;">
<br>
Stuff like that. There were another half dozen or so things there that<br>
I had to tweak to guarantee correct startup/shutdown after installing Maverick,<br>
but I don't remember the others off hand.<br>
<br>
Cheers<br>
<font color="#888888">-ml<br>
</font></blockquote></div><br>