USB Flash Drive Corruption

[RLA113]

Past few months I have been having issues with my USB getting corrupted and locking up Unraid. The server stays on, but is unresponsive. Power button reboot only to have the USB not be detected by the boot loader/BIOS. I have tried different USB ports on my tower, different USB flash drives (both Samsung) and the problem keeps persisting. After the crash I take a flash backup and restore it to the same USB and everything works for a few days to a few weeks, then out of the blue, it gets corrupted again. No major changes recently to my unraid setup.

 

I have no idea what my next steps are for troubleshooting the issue. I am going to try a third USB flash drive when it arrives (Lexar brand 64GB 2.0) and see if that helps. I have also started logging the syslog to a networked computer that I will leave on all the time for now to get the log. Can post that when it happens again. Any help on next steps or what might be causing the issue would be greatly appreciated!

jarvis-diagnostics-20230805-0623.zip

Edited August 5, 2023 by RLA113

[JorgeB]

Are you sure the flash drive is the problem? Enable the syslog server and post that after a crash.

[RLA113]

No, not sure the flash drive is the issue, just the symptom at this point. I have enabled syslog server and will post when I get another crash. Thanks for the reply and help! 

[RLA113]

Got a crash overnight. Was logging the syslog to a remote PC and attached. Let me know if there is anything I can do to help fix this issue.

syslog

[JorgeB]

Not seeing anything that suggests a flash drive problem, I see some older macvaln call traces, make sure you change the docker network to ipvlan, also some miscellaneous segfaults, so good idea to run memtest, issue looks more hardware related.

[RLA113]

Sounds good. I will change over to ipvland instead of macvlan. Running ECC memory so hopefully that is catching an faults.

[electronicbard]

On 8/10/2023 at 3:16 PM, RLA113 said:

Sounds good. I will change over to ipvland instead of macvlan. Running ECC memory so hopefully that is catching an faults.

Were you able to solve your problem? Did switching to ipvlan do the trick? Thanks!

[RLA113]

Sorry for not following up. Took me multiple attempts to fix the issue and a while to troubleshoot and forgot about the thread. Tried changing over to ipvland, but did not fix the issue. I noticed when I could catch the fault quickly, the flash drive would be really hot. So I took a USB extender cable (like the ones that sometimes come with wireless mouse), and placed it away from the back of the case. Now it is staying cool and not having issues. Thinking the exhaust heat was causing the flash drive to heat up to where it was shutting itself down to save itself. My server does not run too hot. Motherboard under 40c and CPU under 60c most of the time, but when the heat from my A2000 GPU started porting out the back it think it was just too much. Since moving it away from straight attached to the motherboard, I have not had the issue. Hope this helps anyone else that might notice their flash drive getting hot.