<br><br><div class="gmail_quote">On Thu, Aug 23, 2012 at 11:33 PM, Joseph Fry <span dir="ltr"><<a href="mailto:joe@thefrys.com" target="_blank">joe@thefrys.com</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
<div class="gmail_quote"><br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">On Thu, 2012-08-23 at 19:20 -0700, Ross Boylan wrote:<br>
> At your implicit suggestion I'll try running memtest and see if it turns<br>
> up anything. The system is a few years old (5?).<br>
><br>
memtest reported no errors after running for over 2 hours and 3 passes.<br>
Ross<br><br></blockquote><div><br></div><div>Did you see if you were able to SSH into the system,</div></div></blockquote><div>I attempted to ssh in, but did not get a connection until the system got itself unstuck. (for both hangs).<br>
<br></div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div class="gmail_quote"><div> if mythweb was still responsive, etc.</div></div></blockquote><div>I don't have mythweb. <br>
</div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div class="gmail_quote"><div> In other words, is it possible that only the video or HID died, making the system appear frozen?</div>
</div></blockquote><div>I don't think so, because ssh and other services were unresponsive too. I'll describe the symptoms in more detail; perhaps you or someone else will understand the role video might have played.<br>
<br></div><div>During the freeze there was a static image on the screen; the display of the time never changed. The mouse pointer was not visible and clicks and typing appeared ineffective. They were not totally lost, because I attempted to switch to a non-graphical virtual terminal (ctl-F2) and when the system came back it did switch to that terminal. Also, the load for the last 5 minutes showed as 60 (!).<br>
</div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div class="gmail_quote">
<div><br></div><div>Reguardless, I am fairly certain, as are others in this thread (well the previous thread before you changed the subject), that your issue is not with Mythtv itself. Perhaps you should try running it outside of the chroot and see if the problem persists?</div>
</div></blockquote><div>I can't do that without upgrading the whole system. When I do it will be to the current stable release, squeeze, while the chroot is testing. It might be an experiment worth making.<br><br>I think for debugging I have to start with the apparent trigger, which is myth. <br>
<br>The "no more queue slots" problem appears to have nothing to do with the video, and is sort of a known bug: <a href="http://code.mythtv.org/trac/ticket/8698">http://code.mythtv.org/trac/ticket/8698</a>. (or the even older 2496) If I read that ticket right it is marked as fixed in 0.24. Unfortunately, I experienced it both in 0.24.2 and 0.25.2. Would it be useful to try to isolate the segment of the video that triggers the problem? Are there some tools I can use to extract chunks from my video? I think I need a sample smaller than the whole recording for useful debugging.<br>
<br>My original thought was that both of the original hangs I experienced had the same root cause. At least on the surface they differed (hang while watching video; hang while transcoding). It's also possible that during the transcoding problem I also had a video problem.<br>
</div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div class="gmail_quote">
<div><br></div><div>Good luck!</div></div></blockquote><div>Thanks; I think I'll need it. <br><br>Swapping could well have a role, as Anthony pointed out. The system can crawl during heavy swapping, and given that the out of memory killer was activated during one of the failures there must have been a lot of swapping before that. I'm not so sure that was the initial problem, however. I did not notice rising memory use before the failure.<br>
<br>Ross<br></div></div>