Geekzone: technology news, blogs, forums
Guest
Welcome Guest.
You haven't logged in yet. If you don't have an account you can register now.


openmedia

3497 posts

Uber Geek
+1 received by user: 902

Trusted

#319482 29-Apr-2025 12:43
Send private message

My MythTV based PVR rig took a hit about 6 months ago during a major power surge and needed a motherboard replacement (​GA-B450M-K V2.0).

 

It initially appeared that the Athon 200GE CPU and RAM we're still functioning correctly, but I'm seeing regular stability issues. In addition I was trying to compile some new DVB debugging tools and the compiler keeps crashing out at random points.

 

I've tried stress-ng and memtest+ and whilst memory appears to be OK after 12+ hours the stress test shows a couple of issues

 

stress-ng --cpu 0 --verify --verbose --timeout 15m

 

stress-ng: debug: [86249] cpu: [86249] started (instance 30 on CPU 1)
stress-ng: debug: [86218] 32 stressors started
stress-ng: debug: [86250] cpu: [86250] started (instance 31 on CPU 1)
stress-ng: fail:  [86223] cpu: drand48 error detected, failed sum
stress-ng: fail:  [86221] cpu: drand48 error detected, failed sum
stress-ng: fail:  [86227] cpu: drand48 error detected, failed sum
stress-ng: fail:  [86245] cpu: drand48 error detected, failed sum
stress-ng: fail:  [86224] cpu: prime error detected, number of primes has been miscalculated
stress-ng: debug: [86223] cpu: [86223] exited (instance 4 on CPU 1)

 

I've now swapped out the CPU for a brand new Ryzen 3400G, but I'm see seeing compiler errors and ongoing errors with stress-ng

 

stress-ng: fail:  [18743] cpu: gray code error detected, sum of gray codes between 0x00000 and 0x10000 miscalculated
stress-ng: fail:  [18734] cpu: prime error detected, number of primes has been miscalculated

 

It doesn't look like a system overheating issue, so I'm going to try a memory swap next. 

 

Meanwhile I'd be interested to hear any other debugging tips. 

 

 





Generally known online as OpenMedia, now working for Red Hat APAC as a Technology Evangelist and Portfolio Architect. Still playing with MythTV and digital media on the side.


Create new topic
openmedia

3497 posts

Uber Geek
+1 received by user: 902

Trusted

  #3368734 29-Apr-2025 12:53
Send private message

Hmm. Looks like it might be memory

 

 

 

stress-ng --vm 4 --fork 8 --timeout 2m --metrics

 


stress-ng: info:  [267135] setting to a 2 mins, 0 secs run per stressor
stress-ng: info:  [267135] dispatching hogs: 4 vm, 8 fork
stress-ng: fail:  [267136] vm: detected 453083136 bit errors while stressing memory
stress-ng: error: [267135] vm: [267136] terminated with an error, exit status=2 (stressor failed)
stress-ng: fail:  [267139] vm: detected 2196476 bit errors while stressing memory
stress-ng: fail:  [267138] vm: detected 5848964 bit errors while stressing memory
stress-ng: error: [267135] vm: [267138] terminated with an error, exit status=2 (stressor failed)
stress-ng: error: [267135] vm: [267139] terminated with an error, exit status=2 (stressor failed)
stress-ng: metrc: [267135] stressor       bogo ops real time  usr time  sys time   bogo ops/s     bogo ops/s CPU used per       RSS Max
stress-ng: metrc: [267135]                           (secs)    (secs)    (secs)   (real time) (usr+sys time) instance (%)          (KB)
stress-ng: metrc: [267135] vm             15546151    120.01    299.96    103.65    129536.73       38518.04        84.08         84604
stress-ng: metrc: [267135] fork             188320    120.00     20.94    517.33      1569.32         349.86        56.07         10484
stress-ng: info:  [267135] skipped: 0
stress-ng: info:  [267135] passed: 9: vm (1) fork (8)
stress-ng: info:  [267135] failed: 3: vm (3)
stress-ng: info:  [267135] metrics untrustworthy: 0
stress-ng: info:  [267135] unsuccessful run completed in 2 mins, 0.02 secs





Generally known online as OpenMedia, now working for Red Hat APAC as a Technology Evangelist and Portfolio Architect. Still playing with MythTV and digital media on the side.




timmmay
20923 posts

Uber Geek
+1 received by user: 5395

Trusted
Lifetime subscriber

  #3368742 29-Apr-2025 13:12
Send private message

Sounds like RAM to me. HCI Memtest sometimes finds RAM problems that Memtest doesn't find.


openmedia

3497 posts

Uber Geek
+1 received by user: 902

Trusted

  #3368748 29-Apr-2025 13:30
Send private message

timmmay:

 

Sounds like RAM to me. HCI Memtest sometimes finds RAM problems that Memtest doesn't find.

 

 

Thanks for the tips, but sadly Windows only. I'll have a look for some more Linux tools.





Generally known online as OpenMedia, now working for Red Hat APAC as a Technology Evangelist and Portfolio Architect. Still playing with MythTV and digital media on the side.




timmmay
20923 posts

Uber Geek
+1 received by user: 5395

Trusted
Lifetime subscriber

  #3368752 29-Apr-2025 13:44
Send private message

If you pay for HCI, and it's not expensive, you can get a bootable USB that doesn't rely on any operating system.

 

In your case it does look a lot like a RAM problem so I'm not sure it is worth the bother. Just replace the ram.


Create new topic








Geekzone Live »

Try automatic live updates from Geekzone directly in your browser, without refreshing the page, with Geekzone Live now.



Are you subscribed to our RSS feed? You can download the latest headlines and summaries from our stories directly to your computer or smartphone by using a feed reader.