For some reason when I build a MythTV frontend and watch a recorded program from the backend, I get a strange pause when I try to jump 30 seconds forward. The little window that pops up in the right corner telling you where you are in the recording will show the current position, and then the length of the recording. The problem is, the length starts out about 4 - 5 minutes longer than it really is and does this weird, sporadic countdown until it gets closer to the actual length and then the jump happens. This can take between 5 - 15 seconds to happen. On my Xbox frontends it works as it should so I don't think the backend or the recordings are bad. I've rebuilt the box a few times and the same thing happens. Here are the specs:

Chaintech 7NIF-4 w/ Duron 1800
512 MB RAM
20 GIG HD
NVIDIA MX4000
KnoppMyth 4 4.1

It does it with the stock 'nv' driver in the XF86Config-4 file and also with the NVIDIA drivers. Has anyone seen this before?

Thanks,
Gonz

Did you ever figure out the solution on this one? I was having the same problem up until today when the time of recording was stretching so far that the jump forward actually turned into a jump back about 2 minutes. At one point the one hour show registered as 10 hours long. It was really odd.

Nope, the problem still exists. I found something on google that said it was corrupt sql tables on the backend. I followed the instructions to clean them up, but it didn't make a difference. I didn't really think it would work since my Xbox frontends work correctly. The problem only shows up on the new PC frontend I've been building. I'm still stumped.

Gonz

Check the DB for errors.

_________________
cesman

When the source is open, the possibilities are endless!

I ran through a doc which had me 'mysqlcheck' all of the tables. Is that what you're referring to? It did find a few errors which I fixed, but the problem remains. A corrupt backend database didn't make sense to me because I have other frontends (all Xbox) that don't have this problem talking to the same backend. Strange stuff.

Gonz

Same here. THere were some problems in the tables, but even after they were repaired the odd problem continues. It is less likely to extend the length of the show, but the length increases and decreases to the original value and in the meantime it loses its ability to skip ahead 30 secs. The first skip in a show may be 30 sec, the next 22, the next 15, the next 3. then not at all.

actualy sounds like a bandwidth problem, you are hitting it because it can't cache enough ahead and keep the feed going. I had a very similare issue with I was on my mesh G network and the spped would drop down to around 12 when streaming vobs.

_________________
-Bushi

That would make sense except in my case the frontend and backend are on the same machine.

Not to mention that I've had 3 different Xbox frontends streaming at once and didn't have the problem. If the PC frontend I built is the only one doing anything on the network I still get the problem. I will try a different NIC just to see what happens. I'm using the built in NIC on the chaintech board so maybe there's something goofy there.

Gonz

Oddity of oddities, it straightened itself out after a reboot. (I did shut off apic in the bios to solve another problem in the process.) I think the mix of fixing the mysql tables and rebooting may have been the key but who knows. However, all of the programs recorded during the problem period are trash. They don't skip ahead properly and throw a lot of odd pixels into the picture every minute or so. They also won't burn to a dvd or video cd. I've simply erased all of them and things seem to be back to normal.

I think I have a little more information to add to this thread.

I have a R4V5 box as my main system. I recently added a laptop (Thinkpad T22, 900Mhz) as another frontend (100Mbit ethernet connection to backend), and it exhibits the same problems as the original poster. It seems that if I do a 30-second skip it pauses for what seems like 30 seconds (making it kinda pointless), and if I do a 10-minute jump it just sits there (perhaps for 10 minutes, but I never waited that long).

I then checked to see how it would behave with videos (not recorded programs). Videos skip and jump exactly as they should. I then went through a whole bunch of my recordings, and discovered that any recording which has already been transcoded jumps and skips properly, too. Only the programs which were recorded and have not been transcoded exhibit this problem.

I am using a PVR 250, and I think all the recording settings have been left at the default, but it's possible I changed something some time ago and no longer remember.

So, last night I set my primary system to automatically transcode after recording...I suspect that will take care of this problem for future recording, but this still warrants further research. If anyone reading this has further troubleshooting requests, I'd be happy to investigate within the limits of my MythTV knowledge.

I have yet some more information to supply on this issue.

I have two independent knoppmyth systems at home now. A R4V5 box (both frontend and backend together), and a R5A12 backend. The laptop I set up for my wife exhibited this skip/jump problem when running as a frontend to either box (using the appropriate version). I also set up a frontend machine to connect to the R5A12 backend. It's config is:
ASUS K8N-E DELUXE (Nforce 3), Gigabit ethernet
Sempron 3000 (underclocked from 1.8Ghz to 1.2Ghz)
GeForce MX4000 AGP. 256MB ram
Sony 8x DVD+RW, 15GB 2.5" hard drive
When I installed R5A12 to this unit, it exhibited the same skip/jump problem. On a whim, I ran the frontend directly from the CD, and it no longer exhibited the problem! I went back and forth a few times, and this is completely reproducable. I then ran the front end off the CD on my wife's laptop (Thinkpad T22, 900Mhz, 100mbit), and it also no longer had any problems with skipping and jumping.

I am very puzzled. It seems that there's some difference in running the front end from the CD vs. running it when it's installed to the HD. It does not seem to be a bandwitdh issue (both 100mbit and Gbit machines ran fine from CD, and had the problem when installed), and doesn't seem to have anything to do with any particular hardware (the ASUS-based machine and the IBM laptop are about as different as possible).

I would really appreciaite it if anyone could give me a clue as to why this makes sense, or if someone can think of something else I can try to narrow down the issue.

Thanx.

Interesting. Try this and let us know.

_________________
cesman

When the source is open, the possibilities are endless!

Thanx for the quick reply; I will try this lilo modification tonight.

But for what it's worth, I may not have been clear about this in my earlier post: we DID see the same problem on a frontend laptop running R4V5, which is an earlier kernel, no? (although I never did try a 'live' frontend on that machine with R4V5. If this doesn't help, I will try that, just to be thorough.)

Regardless, I'll report back later tonight or tomorrow morning with the results.

Thanx again.

Right, different kernel. You're welcome and thank you.

_________________
cesman

When the source is open, the possibilities are endless!