[mythtv-users] Two Persistent Issues - mythbackend & mythweather

Belial Black coroner at darkesthour.com
Mon Mar 29 19:38:22 EST 2004




Overall, mythtv has been a really good experience and I'd like to thank the
developers for their good work. Two things, though, have plagued me since I
began using mythv about five months ago. 

The first should be a very easy fix...I kept meaning to go in myself and weed
out the problem but alas I never can find time to get around to it. In
mythweather, selecting "St. Louis, MO" for location and selecting "i" to save
actually results in selection of Lake Saint Louis, MO (which is a valid but
different entry in the db and is geographically about 100 miles away). The
momemt I save, it goes back to the main weather page and lists info for Lake
Saint Louis. Going back into the configuration screen I find the Lake Saint
Louis entry instead of the St. Louis entry that I selected. This behavior has
been confirmed back to 0.12.

Issue number two is more frustrating and has persisted through three different
linux distributions (I understand that this ultimately makes little difference).
I've swapped distros not in an attempt to resolve the problem but just to try
new things. On RedHat, Gentoo and Knoppmyth, I have consistently had problems
with mythbackend hanging.  Harwdare has remained the same throughout and is as
follows...

AMD Duron 1.3Ghz proc, WinTV Go tv card (bt878), ATI Rage 128 video card (using
GATOS ati.2 drivers), VIA VT8233 sound card

My longest uptime for mythtv is ~30 hours. The uptime generally varies between
12 and 24 hours. Mythbackend then dies leaving a blank screen. I have to switch
to another vtty or log in remotely and restart both mythcbackend and
mythfrontend. Sometimes instead of a black screen, the picture freezes and the
audio gets stuck in a short loop. The key here, I believe, is that fact that
*every time* this has happened the purportedly innocuous "strange error flushing
buffer" is the last entry in mythbackend.log.

It's worth noting that I have reniced every other backend process (including
cron tasks) to ensure that nothing steps on mythtv's toes. That did not help in
any way. I have progressed from 0.12->0.13->0.14 with the same issues. I have
recompiled the kernel to make it as streamlined as possible. No change. I have
changed audio compression settings per suggestions elsewehere in this forum
(turning it off completely actually made mythfrontend start to crash). No luck.
Drive parameters are tuned for the task at hand. 
 
As the mythbackend processes always remains running, I've collected some strace
info if it is of any help in troubleshooting this....

--- SIGRTMIN (Unknown signal 32) @ 0 (0) ---
<... rt_sigsuspend resumed> )           = -1 EINTR (Interrupted system call)
sigreturn()                             = ? (mask now [RTMIN])
rt_sigprocmask(SIG_SETMASK, NULL, [RTMIN], 8) = 0
rt_sigsuspend([] <unfinished ...>


and mythfrontend is left running in this loop...

ioctl(3, FIONREAD, [0])                 = 0
gettimeofday({1080527206, 143808}, NULL) = 0
select(21, [3 5 12 16 20], [], [], {0, 0}) = 1 (in [12], left {0, 0})
read(12, "\0", 1)                       = 1
gettimeofday({1080527206, 144139}, NULL) = 0
nanosleep({0, 100000},  <unfinished ...>




Thanks in advance for any insight into this problem.

*Belial

-------------------------------------------------
This mail sent through IMP: http://horde.org/imp/


More information about the mythtv-users mailing list