[mythtv-users] .25 - "Maximum buffer size exceeded"

[Phil Bridges]

On Wed, May 9, 2012 at 9:09 PM, Ronald Frazier <ron at ronfrazier.net> wrote:
> On Thu, Mar 29, 2012 at 10:32 PM, Phil Bridges <gravityhammer at gmail.com> wrote:
>> I noticed that I had the mysterious "yellow dots" beside some of my
>> recordings this evening.  At second glance, I realized that the two
>> shows I was recording at 9PM had two files.  Not good.  One of the
>> shows was split into two halves, approximately 30 minutes each.  The
>> other though, had only about 11 minutes in it's first half.
>>
>> I looked at the log, and found this:
>>
>>
>> 2012-03-29 21:12:30.800845 E [7151/23462] HDHRStreamHandler
>> ThreadedFileWriter.cpp:216 (Write) -
>> TFW(/mnt/recordings2/1461_20120329210000.mpg:99): Maximum buffer size
>> exceeded.
>>                        file will be truncated, no further writing will be done.
>>                        This generally indicates your disk performance
>>                        is insufficient to deal with the number of on-going
>>                        recordings, or you have a disk failure.
>
>
> I also just got this same error now. I've got plenty of free space on
> the recording drive (29GB, which is about where it always stays). I
> can't find any info about any errors in dmesg or any of the logs.
> smartctl is indicating no errors on any of the drives. I can't find
> any sign of anything actually being wrong. There were only 2 shows
> recording, and I routinely have 4 without issue.
>
> Also, I experienced the same issue as Monkey Pet. After it failed,
> myth continued to report that it was recording in the watch recordings
> screen. The reported file size continued to increase, and when playing
> it, myth continued to report the show growing longer. Now that the
> recording is over, myth is reporting the entire 67 minutes, even
> though it stopped after 29 minutes.
>
> I'm running 0.25 fixes from Sunday morning.
>
> So what could be causing this when I see absolutely no signs of any
> issues other than myth claiming there were issues?
>

Looks like I'm having these problems again.  This time, Myth is trying
to write to a NFS-mounted drive.  There is no notice of anything in my
messages log on either the master backend (which was recording) or the
slave backend (which had the physical drive).

As a side note - I've been running Myth for quite some time, and I
still have recording directories mounted on my master backend that
physically reside on my slave backend.  Should I *not* have these in
the master backend's recording storage group, and only put them in the
slave backend's storage group?