[mythtv-users] possible bug in backend

David Lyons djlyons1 at cox.net
Mon Sep 13 00:10:20 EDT 2004

RAID-0 (striped disk array not fault-tolerant) should not matter since
you are stripping the data however there is no mirror or parity to
rebuild the data if it was deleted. (your wrench in machine).

I have noticed, from time to time, that I may have 60 recorded programs
in Myth front-end. However I may have 70 .nuv (PVR-250) files on my hard

I am not sure why some files do not get deleted from HD when I said
delete the files in Myth. It is something I am trying to figure out
before I submit a possible bug report.



-----Original Message-----
From: mythtv-users-bounces at mythtv.org
[mailto:mythtv-users-bounces at mythtv.org] On Behalf Of Craig Rindy
Sent: Sunday, September 12, 2004 11:43 PM
To: Discussion about mythtv
Subject: Re: [mythtv-users] possible bug in backend

I was wondering about RAID as a possible wrench in the machine.  I'm
using raid-0, and I've run "fsck.ext3 -f /dev/md0" reporting no
problems, so it should act like any other ext3 file system.  Is anyone
else striping with raidtools2?

Cedric Tefft wrote:
> Craig Rindy wrote:
>> Has anyone else run into a problem with their storage needlessly 
>> depleting?  Try recording something and deleting it without
>> the backend, then run:
>> lsof -c mythbackend
>> and see if there are files opened for read access which are deleted.

>> I've been needing to restart the backend everytime I want to reclaim 
>> disk space from deleted recordings for a few weeks now, and it's 
>> getting annoying, so I wanted to see if anyone else has this problem 
>> and submit a bug report if so.
>> I'm using MythTV CVS from yesterday, Debian sid, ext3, and Debian 
>> 2.6.7 kernel with LIRC and HD2000 patches.
> I just checked and I don't have this problem, running CVS from a
> days ago, Fedora Core 2, self-compiled 2.6.7 kernel, ext3.  Maybe it's

> an LVM/raid problem (I'm running neither) or something about the
> kernel?
> - Cedric

More information about the mythtv-users mailing list