Video stuttering: messages about RingBuf, video buffers

[Kantalias]

That drive (ie the new one) will have 4k sectors and if your patitions are not aligned properly it WILL have performance issues.

As far as I understand the issue I do have the partition aligned:

$ sudo fdisk -lu /dev/sdb

Disk /dev/sdb: 2000.4 GB, 2000398934016 bytes
81 heads, 63 sectors/track, 765633 cylinders, total 3907029168 sectors
Units = sectors of 1 * 512 = 512 bytes
Sector size (logical/physical): 512 bytes / 4096 bytes
I/O size (minimum/optimal): 4096 bytes / 4096 bytes
Disk identifier: 0xd8607c12

   Device Boot      Start         End      Blocks   Id  System
/dev/sdb1            2048  3907029167  1953513560   83  Linux

2048*512/4096=256 -> aligned (I think).

Just grasping at straws here but have you tried disabling the WD head parking on that drive?

Yes, I do have the head parking set as least aggressively as possible -- as it can't be entirely disabled -- because I believe the WD20EARS that died leading to the acquisition and installation of this WD20EARX was killed by this very issue. I set the parking timeout to max at boot:

$ cat /etc/rc.local
/sbin/hdparm -S 242 /dev/sdb

exit 0

...it seems you timeouts are much more dramatic than mine. Mine are < 1000ms and yours are well over.

I pasted in the worst of the worst. I see messages ranging from hundreds of milliseconds up to seconds. The hundreds of milliseconds don't seem to hurt anything. Just glancing through the front end log and picking a few of 12,046 lines containing "PrebufferEnoughFrames" since October 4:

Oct 15 20:08:06 mythtv mythfrontend[4247]: N CoreContext mythplayer.cpp:2078 (PrebufferEnoughFrames) Player(7): Waited 102ms for video buffers AAAAAAAAAAAALf
...
Oct 15 20:08:06 mythtv mythfrontend[4247]: N CoreContext mythplayer.cpp:2078 (PrebufferEnoughFrames) Player(7): Waited 203ms for video buffers AAAAAAAAAAAALf
...
Oct 18 21:20:04 mythtv mythfrontend[4247]: N CoreContext mythplayer.cpp:2078 (PrebufferEnoughFrames) Player(e): Waited 101ms for video buffers AAAAAAAfAAALAA
...
Oct 22 18:59:56 mythtv mythfrontend[4247]: N CoreContext mythplayer.cpp:2078 (PrebufferEnoughFrames) Player(h): Waited 1119ms for video buffers AAAfAAAAAAAALf
...
Oct 23 21:39:57 mythtv mythfrontend[4247]: N CoreContext mythplayer.cpp:2078 (PrebufferEnoughFrames) Player(k): Waited 104ms for video buffers AAAAAAAAAAAAAA
...
Oct 27 20:37:57 mythtv mythfrontend[6554]: N CoreContext mythplayer.cpp:2078 (PrebufferEnoughFrames) Player(0): Waited 3463ms for video buffers AALAAAfAAAAfAA

Incidentally, any idea what the parameter in parenthesis after "Player" is?

[nickrout]

As far as I understand the issue I do have the partition aligned:

$ sudo fdisk -lu /dev/sdb

Disk /dev/sdb: 2000.4 GB, 2000398934016 bytes
81 heads, 63 sectors/track, 765633 cylinders, total 3907029168 sectors
Units = sectors of 1 * 512 = 512 bytes
Sector size (logical/physical): 512 bytes / 4096 bytes
I/O size (minimum/optimal): 4096 bytes / 4096 bytes
Disk identifier: 0xd8607c12

   Device Boot      Start         End      Blocks   Id  System
/dev/sdb1            2048  3907029167  1953513560   83  Linux

2048*512/4096=256 -> aligned (I think).

As far as I know starting on 2048 is fine. Whether the end point matters I can't remember

[Kantalias]

As far as I know starting on 2048 is fine. Whether the end point matters I can't remember

It's all about the starting point. As long as the partition begins on a physical sector boundary then every block of the whole partition starts on a physical sector boundary.

The trouble with these dumb Western Digital drives is that they lie about their logical sector size (claiming 512 bytes when it's really 4096) so you have to be careful about the block size of your partition in addition to the alignment.

[nickrout]

It's all about the starting point. As long as the partition begins on a physical sector boundary then every block of the whole partition starts on a physical sector boundary.

The trouble with these dumb Western Digital drives is that they lie about their logical sector size (claiming 512 bytes when it's really 4096) so you have to be careful about the block size of your partition in addition to the alignment.

I am not sure quite what you mean, but what command tests that?

[Kantalias]

I am not sure quite what you mean, but what command tests that?

The logical sector stuff? It's in my post above -- "512/4096" for "logical/physical" from fdisk -lu.

This looks like a good reference for the topic: http://www.ibm.com/developerworks/li...KB-Disksdth-LX.

[frankO]

Could it have anything to do with the video card. I have had the same error for months now. I updated to 12.04 to see if it would fix the problem. It did not. Mine is a front end only. I have other front ends and they work fine. They are all connected by LAN.

Nov 18 17:24:33 darkstar2 mythfrontend[1409]: N CoreContext mythplayer.cpp:2078 (PrebufferEnoughFrames) Player(5): Waited 113ms for video buffers AAAAAADAAAAAAA
Nov 18 17:24:33 darkstar2 mythfrontend[1409]: N CoreContext mythplayer.cpp:2078 (PrebufferEnoughFrames) Player(5): Waited 114ms for video buffers AAALAAAAAADAAA
Nov 18 17:24:33 darkstar2 mythfrontend[1409]: N CoreContext mythplayer.cpp:2078 (PrebufferEnoughFrames) Player(5): Waited 114ms for video buffers AAADAAAAAAAAUA
Nov 18 17:24:34 darkstar2 mythfrontend[1409]: N CoreContext mythplayer.cpp:2078 (PrebufferEnoughFrames) Player(5): Waited 113ms for video buffers AAAAAAAAAAAUDA
Nov 18 17:24:34 darkstar2 mythfrontend[1409]: N CoreContext mythplayer.cpp:2078 (PrebufferEnoughFrames) Player(5): Waited 113ms for video buffers AAAAAAAAAAADAU
Nov 18 17:24:34 darkstar2 mythfrontend[1409]: N CoreContext mythplayer.cpp:2078 (PrebufferEnoughFrames) Player(5): Waited 113ms for video buffers AAAAAAAAAUAAAD
Nov 18 17:24:35 darkstar2 mythfrontend[1409]: N CoreContext mythplayer.cpp:2078 (PrebufferEnoughFrames) Player(5): Waited 117ms for video buffers AAAAAAAAAAAAAA
Nov 18 17:24:35 darkstar2 mythfrontend[1409]: N CoreContext mythplayer.cpp:2078 (PrebufferEnoughFrames) Player(5): Waited 102ms for video buffers AAAAAAAAADAAAA

So my problem is not caused by a disk. This buffer would be in memory I presume, either the cpu or video card ram.

[Kantalias]

Could it have anything to do with the video card.

In general, I don't know if video card issues could cause ringbuf errors in the frontend.

In my case, I don't think the video card is involved -- when I watch a recording (or live TV) while another recording is in progress the incidence of stuttering and accompanying ringbuf errors increases dramatically. Since making a recording only involves getting the data over the network and writing it to disk, I don't think the video card is involved for me.

I have reformatted a WD20EARS to XFS, copied all my recordings over, and swapped it out in place of the WD20EARX. I can't be certain this has had any effect yet since I only did it today. However, I did watch at least an hour of a prior recording while new recording was being made and, separately, hours of live TV with no issue. I will report back in a week or so when I can be certain of the status of my system.

[Kantalias]

No stuttering all week so I think I've solved my problem. Thanks everyone for your help and suggestions.

[bbzzdd]

Glad to hear it's solved on your end Kantalias.

I still have stuttering during commercials which I don't think will ever be solved. The same commercial can play at a different time and day and it will stutter at the exact same place. My problem is probably on the cable co's side.