[mythtv-users] HDHR recording truncated, maximum buffer size exceeded

Tue Jan 1 16:42:33 UTC 2013

Monkey Pet <monkeypet at gmail.com> wrote on Mon, Dec 31, 2012 at 02:52:15PM -0800:
> 
> On Mon, Dec 31, 2012 at 12:11 PM, Joey Morris <rjmorris at nc.rr.com> wrote:
> 
> > I'm running 0.25.3. Twice in the past two weeks I've gotten a
> > truncated recording from my HDHR3. Unfortunately, I didn't check the
> > logs the first time, but last night I did, and the backend log
> > includes an error message saying the maximum buffer size was exceeded.
> > I've pasted that message along with a few that came right before:
> >
> > Dec 30 21:04:02 conquistador mythbackend[9516]: N Expire
> > autoexpire.cpp:640 (SendDeleteMessages) Expiring 15958 MB for 1501 at
> > 2012-12-30T17:08:00 => "NFL Football":"Green Bay Packers at Minnesota
> > Vikings"
> > Dec 30 21:04:36 conquistador mythbackend[9516]: W TFWWrite
> > ThreadedFileWriter.cpp:499 (DiskLoop)
> > TFW(/media/storage1/recordings/1171_20121230202000.mpg:72): write(57904)
> > cnt 997 total 57545296 -- took a long time, 31048 ms
> > Dec 30 21:05:11 conquistador mythbackend[9516]: W TFWWrite
> > ThreadedFileWriter.cpp:499 (DiskLoop)
> > TFW(/media/storage1/recordings/1171_20121230202000.mpg:72): write(57716)
> > cnt 2094 total 120834556 -- took a long time, 33638 ms
> > Dec 30 21:05:20 conquistador mythbackend[9516]: E HDHRStreamHandler
> > ThreadedFileWriter.cpp:216 (Write)
> > TFW(/media/storage1/recordings/1171_20121230202000.mpg:72): Maximum buffer
> > size exceeded.#012#011#011#011file will be truncated, no further writing
> > will be done.#012#011#011#011This generally indicates your disk performance
> > #012#011#011#011is insufficient to deal with the number of on-going
> > #012#011#011#011recordings, or you have a disk failure.
> > Dec 30 21:05:35 conquistador mythbackend[9516]: W TFWWrite
> > ThreadedFileWriter.cpp:499 (DiskLoop)
> > TFW(/media/storage1/recordings/1171_20121230202000.mpg:72): write(57528)
> > cnt 2326 total 134208688 -- took a long time, 22774 ms
> >
> > At the time of the error, Myth was writing that single recording to
> > disk, it was commflagging that recording, and it was possibly deleting
> > another recording. I say "possibly" deleting, because I don't know
> > whether that note means the deletion started immediately, and if so,
> > how long it took to complete. It had just finished (at 21:01)
> > recording another show, and it appears to have recorded that one
> > successfully.
> >
> > A few other details:
> >
> > - I'm recording to a locally attached hard drive.
> > - The drive is dedicated to MythTV recordings. (The OS and database
> >   are on a separate drive.)
> > - The drive is formatted as ext3.
> > - I didn't see anything in the system logs, dmesg, or smart results that
> >   would indicate a drive problem.
> >
> > This behavior is similar to what is described in these two earlier
> > threads, although I didn't see anyone else reporting a file deletion
> > prior to the error:
> >
> > http://www.gossamer-threads.com/lists/mythtv/users/522449
> > http://www.gossamer-threads.com/lists/mythtv/users/524829
> >
> > The last post of that second thread mentions a patch that may help the
> > problem: http://code.mythtv.org/trac/ticket/11061. I think my next
> > step will be to try the patch, but I wonder if it would simply be
> > masking bigger problems in the system, like a failing drive.
> >
> > Do you think this is indicative of a failing hard drive, or do you
> > think the patch really is the appropriate solution? Would I be less
> > likely to experience this problem if I changed the drive to ext4?
> >
> > Thanks for any comments.
>
> This is a long shot, but when I had recording truncation issues, the HDHR
> was dropping off the network and had to continuously ping it to keep it
> alive.  Also why are you still on mythtv 0.25 instead of the 0.26?
> 
> http://www.mythtv.org/wiki/Silicondust_HDHomeRun_Prime#Troubleshooting
> 
> *My recordings are being mysteriously truncated.*
> 
>    - Try increasing the network buffers: sysctl -w net.core.rmem_max=1048576
>    - There is a bug in certain HDHomeRun Prime tuners with the 20120405
>    firmware which leads to it dropping off the network, leading to truncated
>    recordings, broken LiveTV streams, and glitchy recordings. The workaround
>    is to continuously ping the
> HDHR<http://code.mythtv.org/trac/ticket/10414#comment:25>
>    .
>    - Check for packet loss, see the Troubleshooting section on the SiliconDust
>    HDHomeRun <http://www.mythtv.org/wiki/Silicondust_HDHomeRun> page.
> 

(Fixed top posting. See http://www.mythtv.org/wiki/Mailing_List_etiquette#Bottom_post.)

Just to clarify, this isn't a Prime. It's a "regular" hdhomerun.

I tried the low-level check for packet loss, but it didn't produce any
results. The save command didn't produce a series of dots like the
instructions said it would. When I killed it a couple minutes later,
it reported 0 packets received. Watching live TV or recording
something works fine though, so it's obviously transmitting packets.
Are there other ways to troubleshoot hdhomerun network problems? I do
identify the hdhomerun by device ID and not its IP in mythtv-setup. My
hdhomerun is connected to the network via a switch.

Also, I don't understand how networking problems would cause this
error. Why would MythTV complain about exceeding the buffer size when
writing to disk if it's having problems getting data from the tuner?
Wouldn't that result in too little data in the buffer instead of too
much?