| Author |
Message |
|
darkunderlord
|
 Posted: Thu Oct 12, 2006 9:52 pm |
|
|
Joined: Wed Jul 06, 2005 9:27 am
Posts: 28
|
so where are you now? Did you pull your hair out yet? I've gotten this same kind of activity the last couple weeks after several weeks of good recording.
I have a motherboard with:
Athlon XP 2600+
SiS chipset
ATI Radeon 9500
250 GB and 400 GB drives
ATI HDTV Wonder (worked great so far till all these problems)
3 PVR 150's
I'm very upset that this just started happening out of nowhere.
I can't even do a reinstall anymore. I've tried reinstalling to three different drives and still the same issue.
It hangs after the autoconfiguration stuff, where it says something like "Setting Default Volumes". Nothing after that.  
|
|
| Top |
|
 |
|
psicard
|
Posted: Thu Oct 12, 2006 10:18 pm |
|
|
Joined: Mon Nov 28, 2005 10:44 am
Posts: 137
Location:
Austin, TX
|
|
I was having some problems with random crashes as well. Mine seemed to be IRQ related from some of the errors I saw in the logs. I added "noapic" to the "append=" line in lilo.conf. That seems to have cleared things up.
|
|
| Top |
|
|
|
darkunderlord
|
Posted: Fri Oct 13, 2006 7:03 am |
|
|
Joined: Wed Jul 06, 2005 9:27 am
Posts: 28
|
|
Interesting. I was finally able to get through a reinstall by taking out my HDTV Wonder card. i suspected IRQ probs all along, but didn't have any log proof to back it up. The documentation on my motherboard is scarce too so that doesn't help me figure out what's conflicting with what slot.
|
|
| Top |
|
|
|
the_crowbar
|
Posted: Tue Oct 17, 2006 5:25 am |
|
|
Joined: Mon Aug 28, 2006 4:50 pm
Posts: 42
|
|
Been a little while since my last update, but I figured I would post an update.
This box is still no working correctly. I have much more experience with Gentoo than with any other distro, so I installed Gentoo and myth 0.20. With and without the binary nvidia drivers I had frontend crashes. I think I only had one hard losk of the system.
I am currently looking for software to test the various subsystems of the PC. I have run memtest for 24 to 30 hours with no errors. I have run an old hardware test called lucifer. Lucifer detects errors, but does not say what or where those errors come from. I have also tried cpuburn. It runs for some time and then the screen goes blank. Before the screen goes blank I see no errors reported on screen.
If anyone has any suggestions for hardware tests I would love to hear them.
Thanks,
James
|
|
| Top |
|
|
|
darkunderlord
|
Posted: Tue Oct 17, 2006 11:36 am |
|
|
Joined: Wed Jul 06, 2005 9:27 am
Posts: 28
|
|
Yeah I tried adding back in my HDTV wonder after installing and still had issues. I also removed the K20alsa from the startup and also blacklisted the cx88_alsa driver and still have problems. My next step is to go back to the previous version of KnoppMyth.
|
|
| Top |
|
|
|
the_crowbar
|
Posted: Tue Oct 24, 2006 3:02 pm |
|
|
Joined: Mon Aug 28, 2006 4:50 pm
Posts: 42
|
|
I seem to have made some progress on this system. I tried loading Windows XP Pro and MediaPortal (free Media Center type app). I was unable to get recording working while both the PVR-250 and WinTV were installed.
I have an identical Myth system that I upgraded to a 400 GB HD. I took my 250GB drive (that I know is working) and put it in the problem system. For a little while it ran ok and then locked up on me.
I pulled the PVR-250 card and put back both WinTv cards. So far it has been recording and playing back all day without a problem. I don't have sound for one input so my friend is headed to BestBuy to buy an USB sound card.
Tomorrow I should be able to have everything up and working. I have it set to record all night long.
Thanks,
James
|
|
| Top |
|
|
|
darkunderlord
|
Posted: Tue Oct 24, 2006 5:00 pm |
|
|
Joined: Wed Jul 06, 2005 9:27 am
Posts: 28
|
|
Crow, you ever play RTCW?
I ended up using a different machine. No problems now. I got tired of trying things.
|
|
| Top |
|
|
|
tjc
|
Posted: Tue Oct 24, 2006 8:18 pm |
|
|
Joined: Thu Mar 25, 2004 11:00 am
Posts: 9551
Location:
Arlington, MA
|
|
Have you tried derating the RAM timings? At this point I doubt that's it, but it's easy to try. I think you've already tried adding "noapic noioapic" to the lilo boot command, right?
Other than that you're probably looking at some ugly interupt or DMA conflict.
|
|
| Top |
|
|
|
randomhtpcguy
|
Posted: Fri Oct 27, 2006 7:48 am |
|
|
Joined: Mon Nov 07, 2005 10:09 am
Posts: 153
|
|
I have had similar but more infrequent random lockups. Have you tried mbmon and smartmontools. I found that one of my hard drives runs at 100C while the other runs at 50C. I don't know if the northbridge temp is an option in mbmon, but i guess the feel test works just as well.
Since you did such good detective work, I hope you can go the next step and submit a bug report if its more than just defective/broken hardware. I think a lot of people are having issues with the pvr cards and ivtv with r5d1, but i don't know if its the same as you. i am waiting to upgrade until they are worked out in a release.
|
|
| Top |
|
|
|
borgednow
|
Posted: Fri Oct 27, 2006 10:04 am |
|
|
Joined: Wed Aug 23, 2006 3:07 pm
Posts: 339
|
the_crowbar wrote: Thanks for the reply tjc!
The system is powered by a UPS (APC 1100 VA or some such, plenty of power for the myth box).
I am about to have the system back on a workbench, but I don't know what may be the most helpful in regard to troubleshooting.
I am thinking of switching back to the stock kernel (SMP) because that is what I run on my box. I will however loose the linux-wlan-ng drivers and need to recompile them. Of course I can't imagine why a SMP kernel would be more stabel than a non-smp kernel. I used the same source and same config except for the SMP change.
It seems the box crashes more frequently running the frontend (watching something) than just recording shows with the backend.
I will post again once I can test with the stock KnoppMyth kernel.
Thanks,
James
Just having a UPS doesn't mean the power is clean.
I've been having weird problems for the last few months where my primary box would shut down randomly anywhere from a few times a day to once every few weeks.
I finally unplugged the machine from the ups and just straight into a surge protector instead. So far no more shutdowns.
Just so you know for sure, remove the ups from the equation.
|
|
| Top |
|
|
|
the_crowbar
|
Posted: Wed Nov 01, 2006 11:57 am |
|
|
Joined: Mon Aug 28, 2006 4:50 pm
Posts: 42
|
|
Updated status:
System has been running ok. Rare lockups with just the WinTV cards (no PVR-250).
borgednow: I have experienced the crashes both connected to mains power at my office and connected to the UPS. I think I can rule out the UPS as a problem.
I will now try to slow down my RAM settings, it is just hard when the crashes are very rare. The system may run for days without a crash.
I do need a little help with the dual audio cards changing their /dev/dsp entries after each boot, but I posted another thread about that.
Thanks,
James
|
|
| Top |
|
|
|
bruce_s01
|
Posted: Wed Nov 01, 2006 6:33 pm |
|
|
Joined: Tue Aug 08, 2006 7:08 pm
Posts: 561
Location:
UK
|
|
As a suggestion, check your PSU and DC power draw, it's possible that a particular powerline is marginal, particularly if you are using two drives.
Do you have a higher rated PSU available to try out?
Bruce S.
_________________
Updated 2019/10/26: AthlonII X2 265 Gigabyte GA-970A-DS3P
16Gb PC 1866 DDR3, 500GB+2TB+4TB SATA HDD,
SATA DVD-RW Asus DRW-24D5MT , NVIDIA GeForce GT1080
Hauppauage Nova-T 500, Nova-T LinHes R8.6.1
|
|
| Top |
|
|
|
Wizzard~Of~Ozz
|
Posted: Wed Nov 01, 2006 6:55 pm |
|
|
Joined: Wed Apr 12, 2006 9:37 am
Posts: 30
|
Quote: After applying a generous amount of Artic Silver the heatsink was attached to the MB.
As with any heatsink compound, more is not better. after applying the compound you should be able to see the writing on the chip. it is meant to fill microscopic grooves in the chip and heatsink, direct contact is still better.
As for your issue, Often random hardlocks are memory, as you've tested this, it would be unlikely.
Some things to try, after you install both cards, try selecting "reset configuration data" in the BIOS (should be under PCI/pnp). Also, while you're in there check the temp of the CPU, also watch the voltage for the CPU for spikes (or dips), you can try increasing the voltage by .025v to increase stability of the CPU if your BIOS allows.
Also wondering what PSU you have in this box? should be at least 400W Name brand (Antec, Enlight or the like). Often some of the rails are underpowered (12v is typically the culprit).. On the side of the PSU it should list Voltage/currents in a chart.. your 12v rail(or rails) should be 20A or higher..
_________________
AthlonXP 2500/768Mb DDR
AbitNF7 (onboard Audio),Geforce 5200
NFS share for 1.5tb (3 x 500Gb) w/ XFS under FC7.
ATI HDTv Wonder, Hauppauge 150MCE
ThermalTake Bach case
|
|
| Top |
|
|
|
the_crowbar
|
Posted: Thu Nov 02, 2006 8:10 am |
|
|
Joined: Mon Aug 28, 2006 4:50 pm
Posts: 42
|
Thanks for all the suggestions!
After adjusting the memory timings in the bios, I received my first bit of information in a log file. Normally the system hard locks. This time I got some DMA errors related to the PVR-250 card.
Code: Nov 1 16:21:23 mythtv kernel: ivtv0 warning: ENC: (0) DMA Error 0x0000000b 00000001
Nov 1 16:21:23 mythtv kernel: ivtv0 warning: offset 0 -> 128
Nov 1 16:21:24 mythtv kernel: ivtv0 warning: offset 128 -> 100
Nov 1 16:23:34 mythtv kernel: ivtv0 warning: ENC: (0) DMA Error 0x0000000b 00000005
Nov 1 16:23:34 mythtv kernel: ivtv0 warning: offset 100 -> 128
Nov 1 16:23:34 mythtv kernel: ivtv0 warning: offset 128 -> 0
Nov 1 16:30:30 mythtv kernel: ivtv0 warning: ENC: (0) DMA Error 0x0000000b 00020001
Nov 1 16:30:30 mythtv kernel: ivtv0 warning: offset 0 -> 128
Nov 1 16:30:30 mythtv kernel: ivtv0 warning: offset 128 -> 116
Nov 1 16:31:37 mythtv kernel: ivtv0 warning: ENC: REG_DMAXFER 2 wait failed
Nov 1 16:31:37 mythtv kernel: ivtv0 warning: ENC: (0) DMA Error 0x00000001 00000001
Nov 1 16:31:37 mythtv kernel: ivtv0 warning: ENC: REG_DMAXFER 2 wait failed
Nov 1 16:31:37 mythtv kernel: ivtv0 warning: ENC: (1) DMA Error 0x00000001 00000001
Nov 1 16:31:37 mythtv kernel: ivtv0 warning: ENC: REG_DMAXFER 2 wait failed
Nov 1 16:31:37 mythtv kernel: ivtv0 warning: ENC: (2) DMA Error 0x00000001 00000001
Nov 1 16:31:38 mythtv kernel: ivtv0 warning: ENC: REG_DMAXFER 2 wait failed
Nov 1 16:31:38 mythtv kernel: ivtv0 warning: ENC: (3) DMA Error 0x00000001 00000001
Nov 1 16:31:38 mythtv kernel: ivtv0 warning: ENC: REDO DMA took too many tries.
Nov 1 16:31:38 mythtv kernel: ivtv0 warning: Error Encoder DMA
Nov 1 16:54:12 mythtv kernel: ACPI: PCI interrupt for device 0000:00:10.1 disabled
Nov 1 16:54:12 mythtv kernel: bttv0: timeout: drop=37 irq=500643/1454930, risc=02da807c, bits: HSYNC OFLOW FDSR
After a Google search I saw a thread about this exact error on the ivtv mailing lists. It seems there are still bugs in the ivtv drivers ( I am using 0.7.1).
bruce_s01: The system is powered by an Antec 350 watt PSU. I think it should be more than sufficient. I have an identical system at home powered by a 280 watt PSU and it has been running fine under heavier load (HD Video) for months. I don't have another spare PSU to try, but I can pull one from another computer if need be.
Wizzard~Of~Ozz: I know too much thermal compound can be bad, but I used Artic Silver compound and don't think I used too much. The CPU heatsink never gets warm to the touch. It is only an Athlon 64 3000+. The Athlon64 chips run pretty cool. Right now I am trying to get sensors up an running on KM, so I can read the sensor chip under the heatsink.
Thanks,
James
|
|
| Top |
|
|
|
the_crowbar
|
Posted: Thu Nov 02, 2006 11:44 am |
|
|
Joined: Mon Aug 28, 2006 4:50 pm
Posts: 42
|
|
After fiddling with the PCI latency_timer settings I may have made a little progress. In the thread on the ivtv mailing list they talked about PCI latency.
The first thing I tried was setting the latency in the BIOS to a higher value. It was on 64 and I bumped it up to 128. When I tried to watch through the PVR-250 card I got an immediate hard lock.
The next idea was increasing the latency_timer of the PCI devices in the computer to a value higher than the bios value. (The WinTV and PVR-250 by default have a latency of 128.) This resulted in a hard lock.
Next idea was to set the latency in the BIOS below what the PVR-250 was set to. I changed the BIOS to 32 and left the WinTv and PVR-250 at their default of 128.
So far I have been recording on both cards for an hour. I do have an issue with getting the backend status in MythWeb. I am not sure if the frequent resets may have corrupted the fs somehow. I would think that a system reset would only corrupt files being written to, but who knows.
I'll post again as I find out more info.
Thanks,
James
|
|
| Top |
|
|
View unanswered posts
View active topics
Print view
Previous topic Next topic 
LinHES Home
Board index
FAQ
Search
Login 
Register
