[mythtv-users] recorded TV playback jumping around

Sam Hooker mythtv at noiseplant.com
Sun Sep 21 05:49:29 UTC 2008


-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1


Hi folks,

I've been a happy mythtv user for two or three years, now. Currently, I
have 0.21-192 running on Fedora 9 (P4/2.8/512MB) as my primary backend
and primary frontend machine. When running 0.20 on Fedora 6 previously,
playback of recorded TV (analog cable) was pretty flawless on this same
chassis (with different disks; see below). Since upgrading to Fedora 9
and myth 0.21, I've noticed two oddities:

1) playback of recorded TV jumps around at random times, at
seemingly-random intervals, and in both directions (forward an
backward); playback continues from the "landing point", unless of course
that's after EOF

2) 'top' shows between 18% and 28% iowait at all times, even when the
system isn't doing anything (recording/playback/commflag/RAID resync/etc.)

#1 is my primary concern, as it's really annoying. Especially to my wife
who, despite not being a technophile, loves Myth. This is starting to
modulate her opinion. ;-)

What else changed, other than software, in the upgrade? The disks:
previously, I had the video stored on a JFS logical volume (LVM2)
sitting atop three IDE disks, which was migrated to an ESATA disk before
that pooped out. (Both of those arrangements performed well, though.)

Now, everything (OS and video) sits on a pair of 500GB SATA2 disks in
multiple software RAID1s connected to a PCI SATA1 card. (The video
itself still sits on its own RAID1'ed partition: JFS over LVM2.) Taking
all this into account, #2 represents a not-unforeseeable state of
affairs...layers upon layers, and all that...)[1]

When the jumps happen, the OSD pops up with the message you'd expect if
you had actually told mythfrontend to jump ahead/back. Neither 'dmesg'
nor /var/log/mythtv/mythbackend.log shows anything telling (disk errors
included). In a moment of weird paranoia, I firewalled off tcp/6546,
thinking there was a vanishingly-small chance that something else on my
network was sending spurious commands to MythFrontend Network Control:
the problem persisted even after that. Apparently, I just have PoulterMyth.

I'm currently watching tcp/6546 on eth0 and lo, and running 'irw' to see
if there's a new IR device chattering away in my living room. Oh, and
the system is reasonably up-to-date (there are two updates waiting:
ksysguardd and xine-lib-moles).

Has anyone else seen anything like this?


Cheers,

- -sth

[1] I realize that the disk situation is suboptimal from a performance
standpoint, but I only have two analog tuners, and the throughput should
be good enough for at least two recordings and a playback or two. After
losing the majority of my media in the aforementioned disk failure, I'm
willing to put up with _some_ penalties for using RAID1. (This presumes
that the "random jumping" isn't somehow related to I/O starvation, which
seems unlikely to me.)

sam hooker|mythtv at noiseplant.com|http://www.noiseplant.com
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.8 (Darwin)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org

iEYEARECAAYFAkjV4GkACgkQX8KByLv3aQ3rhACffXxQu5b1Y6B7ibxoe3n0VO91
Bp4AoPVC0e5Ia4Ths/rPhZ8VHVW7XlkR
=53Cy
-----END PGP SIGNATURE-----


More information about the mythtv-users mailing list