Player lockups on 1.03?

Can anyone duplicate this behavior?

My friend Tod has been having his player lock up regularly while it's playing on his desk. For a while, I assumed it was just the known lockup bug in 2.0 beta 7.

But he downgraded to 1.03 and still gets occasional lockups, although less frequently. Perhaps there are two separate things causing lockups.

We think it might be USB-related because he usually has the USB cable plugged in, and he had one player lockup which also locked up his USB "Strategic Commander" input device. He's going to try leaving his USB cable unplugged for a while to see if that stops the lockups.

Today, with 1.03, he said the music stopped playing and the front-panel stopped responding, but the visual kept scrolling (although without musical input). Does this last one sound like a disk drive or cable problem? He hasn't had the unit freeze-on-boot as far as I know, reboots always seem to fix the trouble for him.

I explained to him that with 1.03, there should never be a lockup. I was able to play my player for literally MONTHS without a lockup on 1.03. He said he didn't believe me and that there's got to be other people with the same problem. Anyone else have this trouble besides the DOA/Drive-cable people?

My next plan of attack will be the swapping of the disk position to the second head on the drive cable. Anyone else agree that's the correct next step?

Okay, Tod's player locked up again, and the following new information is hereby offered:

- He is currently running plain-jane 1.03-consumer, with no special kernels or other software installed.

- The lockup does not appear to be a specific song. He has had the lockup happen on specific, small, playlists which (in the past) have played through beginning to end. It seems to happen randomly at any time, and does not seem to be song-related or playlist-related.

- The lockup happened when there was nothing plugged into the player except the RCA audio cables and the power. No other cables were plugged in. It was just sitting on his desk playing music. He is using the AC adapter that shipped with the player.

- I, personally myself, moved the hard disk to the end connector on the drive cable in hopes that it was the bad-connector problem. I did it correctly, by actually moving the drive itself to the other spot on the tray so that the cable didn't have to be twisted in a strange way. This did not change the behavior.

- His player passes a ctrl-T memory test.

- The lockup manifests itself as follows: Music stops playing. User interface stops responding to button presses from either the front panel or the remote. Visuals will continue to animate in "free-wheel" mode as if they were not getting sound input.

- The lockups are very intermittent. Perhaps once a day on average.

- He did not have his hard disk indicator icon activated, so we do not know the state of the hard disks. I have asked him to activate the indicator and report its state on the next lockup.

Anyone? Anyone?

I just got my players back yesterday and had the one running for maybe 20 minutes when it locked up playing music. I was running 1.03 at the time with a ethernet connection going but this time the front buttons still worked and changing playlists brought the music back. The first time it happened I had to pull the power and it started back up where it left off. Both times the visuals kept going.

It has only happened twice, on different songs and different playlists. Last night I also had Emplode open but I can't remember the other time if it was. I would assume since David worked on it for upgrading the drive that it wouldn't be the cable. I also do not have the hard disc indicator selected. I upgraded it last night to running 2.0 beta 11 and 1.99 hijack.

Thinking back I believe the first time I was running 1.03 consumer and the second time it happened I was running 1.03 developer. I would be willing to go back to running 1.03 consumer to see if it keeps happening to me. And if you could tell me how to log any of the events, I will.

So he isn't the only one who has had lock ups.

Get a serial log of the player when it actually locks up.

Should he have the developer build installed to get the serial log when it locks up?

I think consumer outputs major errors.

Okay, he's going to hook up to the serial port today and hope it locks.

Should we also be getting a new cable, or is the drive-position-swap enough to rule that out as a possible cause?

Okay, he got a lock with the serial console up.

There were no messages displayed on the serial port at lockup time. The serial port was working because he could do "n" and "p" and they worked. And then it put up the usual "DHCP Discover" messages, then it locked.

Also note that he put the disk indicator icon up and it was stuck on with the gray line in the center of the icon at lockup time.

Probably not related, but when he boots up, it reports his hard disk multiple times. Is this normal?:

Probing primary interface...
hda: FUJITSU MHM2100AT, ATA DISK drive
hda: FUJITSU MHM2100AT, ATA DISK drive
hda: FUJITSU MHM2100AT, ATA DISK drive
hda: FUJITSU MHM2100AT, ATA DISK drive
hda: FUJITSU MHM2100AT, ATA DISK drive
hda: FUJITSU MHM2100AT, ATA DISK drive
ide0 at 0x000-0x007,0x038 on irq 6
hda: FUJITSU MHM2100AT, 9590MB w/2048kB Cache, CHS=19485/16/63

Tony, I got this same multiple reporting of the hard disk when it would lock up (for my disk upgrade problem). It seemed rather random, but it would always do either that (and freeze on the boot logo) or give the "Hard Disk Not Found" error.

Ended up being a bad cable for me.

He performed a full FSCK from the shell prompt (per the instructions in the FAQ), and he did not get any errors, his disk checked fine.

He did notice the following at the shell prompt, though:

E !! poll_engine.cpp 41: Failed to poll descriptors.
E !! poll_thread.cpp 46: Failed to poll 0xc000000a.

Is this normal for 1.03?

Maybe we should just ask David for a new cable, to be absolutely sure.

Tony,

I've reported this last week on this board, but maybe you've missed it (doubtful, more likely you forgot you've read it ).

I too have gotten lockups with 1.03. On my unit (a Rio60gig), only a plain developer image was installed. No kernel hacks.

The player has locked up on me on a ratio of once in 1-2 days. I wasn't able to reproduce the problem on demand. It seemed very random. It wasn't related to a certain file whatsoever.
The unit locked up on me in both the car and at home. Each time it required unplugging the power and having it reboot cold.
I've upgraded to 2.0 beta 11 last week and have not experienced the problem since. I hope it's solved for me now.

I doubt it has something to do with USB because I haven't used the USB connector on the unit even once.

I might be a bad cable because I've ordered 3 more 60 gig players and I guess they all came from the same batch (serial #'s that weren't to far apart from each other). All other 3 units required a new HD cable because the HD's simply wouldn't mount. I thought I was lucky to have a least ONE unit work correctly without having to open it and swap the HD cable.
Then again, it might not have been the cable at all. Because as said before, upgrading to b11 solved it for me (up til now anyway - touch wood)

I'm sorry I can't give you a solution or pinpoint the cause of the problem.

Did you get a new cable into the player that was locking up? Did it solve the problem?

No, I didn't. I was actually glad I could leave one player as-was. So now I've still got one player of which the seals aren't broken. But if the bug returns it will be the first thing I'll try.
Sorry!

Yes, it's normal - it's reprobing to see if a second drive turns up. It gives up after several tries or if it finds a second drive.

Hugo

Here's something that might be of interest.

He installed 1.03 developer, and it hasn't locked up for a day or two. The lockups seem to have only been under 1.03 consumer.

This totally doesn't sound right to me. One would think that the two platforms would be equally stable, or if there were a difference, that developer would be the one to crash more often.

Bad download of 1.03Consumer? Or would that more likely result in the install not even working?

Well, he got 1.03 consumer off of the CD that came with the tuner. That still doesn't rule out file corruption, but reduces its likelihood.

I gotta admit, it feels kinda funny throwing ideas out to the FAQ Master! Other than a "we snuck in the room and sabotaged your buddy's RioCar while you were out of the room" I don't think many of us here could offer any suggestions that you havn't already thought of!

Despite rumors to the contrary, I am human and therefore do require help once in a while.:)

Here's an idea:

What if it only happens on Mk2a players (not Mk2), and only if Consumer (not Developer)? There's one thing that those datapoints have in common: Memory footprint. The 2a's have a larger memory size than the 2's, and Consumer uses more memory for cache than Developer. Perhaps there's something about the cache code that fails intermittently with that particular combination?

In the past I've had a few reports of the upgrade file on the CD being corrupt - as far as I am aware, they always resulted in the upgrade failing, but maybe this one got through.

Get him to try installing a fresh download of 1.03 consumer.

Okay, disregard the last stuff. He had it lock with 1.03 developer today.

What are we down to? A bad cable? A bad drive?

Well, 4 days on b11dev and no lock-ups. Go figure.

Hmm. Possibly corrupted copies of 1.03 on the CD-roms?

Anyway, glad you haven't had any probs with B11. All's well that ends well, eh?