[mythtvnz] Recording dropouts and disk performance...

[Rob Connolly]

On Sun, Feb 17, 2013 at 03:42:33AM +1300, Chris Bannister wrote:
> On Sat, Feb 16, 2013 at 03:32:18PM +1300, Rob Connolly wrote:
> >  
> > Hi All,
> > 
> > Over the past couple of weeks I've had a few recordings drop out after
> > several minutes. Today I finally got around to looking in the log files
> > for one of the affected recordings:
> > 
> > 2013-02-14 21:38:59.782815 E [1609/24862] DVBRead ThreadedFileWriter.cpp:217 (Write) - TFW(/mnt/recordings/1001_20130214083000.mpg:110): 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.
> > 
> > My disk setup is 2x2TB drives set up using software RAID 1 and LVM, with
> > the system and database partitions being hosted on the same drives as
> > the recordings. This setup has worked fine until recently - I did have
> > one drive die a while back, but I returned it under warranty.
> 
> Have you had a look at hdparm?
> i.e hparm -tT /dev/sdx
> (where x is the letter of your device)
> 
> Is dma turned on for all drives?
> 
> I have no experience with RAID, so don't know how useful hdparm is in a
> RAID setup.
> 

Here is the hdparm output for both disks:

/dev/sda:
 Timing cached reads:   3074 MB in  2.00 seconds = 1536.81 MB/sec
 Timing buffered disk reads: 332 MB in  3.00 seconds = 110.54 MB/sec

/dev/sdb:
 Timing cached reads:   3052 MB in  2.00 seconds = 1526.28 MB/sec
 Timing buffered disk reads: 342 MB in  3.01 seconds = 113.48 MB/sec

/dev/sda:
 multcount     = 16 (on)
 IO_support    =  1 (32-bit)
 readonly      =  0 (off)
 readahead     = 256 (on)
 geometry      = 243201/255/63, sectors = 3907029168, start = 0

/dev/sdb:
 multcount     = 16 (on)
 IO_support    =  1 (32-bit)
 readonly      =  0 (off)
 readahead     = 256 (on)
 geometry      = 243201/255/63, sectors = 3907029168, start = 0

However, when I try to probe for DMA support I get the following, on
both disks:

/dev/sdb:
 HDIO_GET_DMA failed: Inappropriate ioctl for device

I'm pretty sure something is not right here, since DMA is a pretty basic
feature!

Cheers,

Rob