Any suggestions before I stop using SageTV?

[Savage1701]

Oh, are you network controllers' settings correct with things like scaling, packet size, etc.?

Again, that should not corrupt a strict playback, but I'm grasping at straws here.

I figure since you do this sort of thing you have tested your GbE line quality, loose punchdowns, make sure you have good switches/routers, patch panels, etc.

That's about it on my guesses. I really hope you get it fixed.

[Dave47]

Originally Posted by Rezolution

If you're 100% certain it's not the PC, it's possible that corruption is occurring at the network level. Maybe there is a bad NIC/Switch/Router/Cable somewhere in your network that is causing bad data to be sent back to the sage software running on your main PC.

I think I had this network issue once upon a time. My setup has been stable ever since I upgraded to a 24 port gigabit switch that covers the entire house, pulled brand new CAT6 wire, and rewired every wall jack in the house. My house used to be a mix of 5 and 8 port gigabit switches connected to various CAT5e cables and possibly suspect wall jacks.

David

[Savage1701]

Stupid idea #101:

How about returning to the days of yesteryear and trying a .mkv off of a drive that is set to read-only? Say, a NAS/share from another PC?

If the drive is write-protected, then the data can't become corrupted, correct? I'm thinking archive-type drives that might have old scans of financial data or spreadsheets on them, where you want people to be able to look at them but never change them.

Could you try the same thing with your setup? Another computer on the network set to have a share set to "read only"?

I suppose the ultimate reducto ad absurdum would be to re-burn a .mkv file to a dvd and see if the corruption occurred using Sage as a DVD player.

Or a write-protected thumb drive, I don't know.

If you got corruption after several days of viewing, you know it CAN'T be the thumb/optical/share drive because the dvd has been burned and closed and is truly read-only, or the drive share is set to read only or the thumb drive has a write-protect feature on it.

At least that would point you towards some other aspect of file corruption and away from SATA cables, hard drive issues where the .mkv resides, etc.

I don't know where you would look next, but you might know a few areas you could stop looking.

I'd still be suspicious of USB to SATA bridges in external enclosures. I had a couple that could not handle much of what Sage threw at them or tried to read off them, even with the cache enabled.

[Savage1701]

Originally Posted by Dave47

I think I had this network issue once upon a time. My setup has been stable ever since I upgraded to a 24 port gigabit switch that covers the entire house, pulled brand new CAT6 wire, and rewired every wall jack in the house. My house used to be a mix of 5 and 8 port gigabit switches connected to various CAT5e cables and possibly suspect wall jacks.

David

+1 on that front as well.

I'm sure others mentioned this many times, but I am also wondering - bad sectors/re-allocation of blocks/error recovery compounding the issue on itself? This thread is getting long enough I can't remember what everyone else has posted.

I'm just not wrapping my head around how it's corrupting what should be a read-only situation.

[speck55]

Originally Posted by panteragstk

You don't have to worry about a key. Just use the server in trial mode. You have 21 days to figure out the issue.

With all the hard drive changes I'm hoping this is just some sort of software bug. It could be a simple driver issue. You wouldn't believe how many issues a bad NIC driver.

I had some buggy behavior like this, and I was chasing my tail until it dawned upon me to check the Windows Event Log and found constant controller resets on my ESata controller by the driver. Replaced ESata controller, and no more "XXXX initiated a reset on device Harddisk/DR11" or similar. Look for entries with a source of "Disk" in the System event log. (If you've already tried this, ignore me. ) - but it was something that was living heck to chase down, and I was beginning to think I needed to quit sage. What I needed to quit was buying too-cheap-to-be-true controllers off ebay.

[elefante]

Also I have run prime95, memtest86 and OCCT test for the CPU and RAM, always pass without issue. The video is brand new to replace an old power hungry 4850 (a client had an insurance claim and let me keep the rest of the system mobo got fried).


Are you saying that you are using hardware from a PC that had a fried mobo?

Also this sounds like you are having storage problems in the drive where you are hosting the MKV. It is either the SATA cable (it happens more than you think), the connector to Mobo, or the hard drive.

You say the CPU/RAM are new, however you point to a USB 3 controller for BD, which I am assuming you are adding a USB3 card?

Pls specify the AMD mobo.

Also, did you update firmware on mobo to a compatible one for the proc.

Is the RAM DDR2 or DDR3?

Also w/ AMD mobo's I have found running the drives in IDE mode works best, because the AMD hard drive SB drivers are atrocious and they dont fully support SATA NCQ so turning on AHCI is just asking for trouble.

I also had a realtek onboard Enet port go bad and the studdering issue you spoke about was on my system. I replace w/ Intel PCIe and it went away.

However if you are saying the actual MKV file is corrupted, take an MD5 hash (or the like), move it to another drive and take the hash again. They should be identical.

When you generate the MKV, etc take an MD5 hash and keep it (I always do). This way if you are seeing corruption you know the file has been modified (corrupted) at a later time. I use MKV for years because I like DD in the container, and I have had zero problems w/ them and I just installed ffdshow tryouts and MKV splitter (Haali). I use 32bit Win7 tho.

If it is corrupted on the new drive, then with almost certainty you have a data instability in the drive path (cable, SATA, drive, etc).

If the hash is different, you may be seeing ECC issue on the drive (silent corruption) and it is most likely the drive, but still could be any of the other components.

Also you can get the SMART data from the drive to see if there are mortality issues in play.

I have a customer w/ over 10k drives in production and they have over a dozen checksum corruptions every day. This happens, but we looked more closely at the issue at we found that the SAS extender cables were exceeding their bend radius and that was causing periodic corruptions.

[Savage1701]

Originally Posted by speck55

I had some buggy behavior like this, and I was chasing my tail until it dawned upon me to check the Windows Event Log and found constant controller resets on my ESata controller by the driver. Replaced ESata controller, and no more "XXXX initiated a reset on device Harddisk/DR11" or similar. Look for entries with a source of "Disk" in the System event log. (If you've already tried this, ignore me. ) - but it was something that was living heck to chase down, and I was beginning to think I needed to quit sage. What I needed to quit was buying too-cheap-to-be-true controllers off ebay.

+1 on the eSATA controller as well - I had a SOHO system used for something completely different that would periodically drop an eSATA port connected to an external drive/enclosure.

Never did eSATA with Sage, but it sure raised heck with the SOHO system when it happened, which was a few times a week. Took me a little while to catch it because it only seemed to happen during periods of sustained drive activity.

I'm with Speck - I'd check the logs when you are getting the corruption.

[equalize]

If this were me, I'd
1. Check System Event logs
2. If running an anti-virus, see if you can exclude .mkv files
3. Check if the hard dive manufacturer has a utility to check the health of your drives.
4. Check and/or replace hard drive cables
5. Run Process Monitor and focus on File System Activity (technet.microsoft.com/en-us/sysinternals/bb896645)