I keep getting issues with docker. Docker crashing randomly until the full docker service get KO (won't start/stop, image corrupted etc....).

 

I had to re-create the docker image 3 weeks ago due to this kind of issue

Last week I couldn't event start docker without the system getting stuck

 

I restarted docker on Saturday.

Two days ago Influxdb randomly stopped. I restarted it and it's now working since.

This night I had CA Backup/restore scheduled and it seems to have cause an issue with Bazarr. Bazarr wasn't working (but started), constant `OSError: [Errno 5] I/O error`in bazarr's log. And in the server's syslog I see continous.

BTRFS warning (device loop3): csum failed root 1069 ino 8067 off 24199168 csum 0x4bd3e39b expected csum 0xc589050e mirror 1

The warning stopped when I stopped Bazarr

My cache/ssd is xfs formated so the warning is related to the docker image. I ran a scrub operation that detected 7 error but could'nt correct them:

Error summary:    csum=7
  Corrected:      0
  Uncorrectable:  7
  Unverified:     0

And that just the start, I'm pretty sure that by the end of the week the full docker stat won't be working.

 

I'm not sure what could be the source of all this issues.

I plan to reformat the ssd/cache drive but I'm not sure that will definitely fix the problem. Does anybody is familiar with this kind of issues or could give me a hint to what could be the source?

 

 

syslog_2020-12-22.log

Checksum errors suggest a hardware problem, like bad RAM, you are also getting multiple segfaults, which suggest the same, since you didn't post the diags we can't see if the RAM is overclocked, check here and if not run memtest.

Here are the full log.

The hardware hasn't change since I built it one year and a half ago. Except for the ssd that was dead and replace for a nvme drive in June.

The issues does seems to have appeared during this year, so I don't think it's related to an overclock of the ram. I'm not able to run a memtest since I don't have a graphic card.

 

intersect-diagnostics-20201222-1344.zip

6 minutes ago, seneo said:

it's related to an overclock of the ram

It's not, possibly just bad RAM.

 

6 minutes ago, seneo said:

I'm not able to run a memtest since I don't have a graphic card.

Recommend you get one.

I've just ordered a graphic card and new memory sticks. Let's hope that the source of all the issues. I was already thinking of ram but with you input I'm more confident that the main issue.

Thanks for your help.

One week since I put the new ram and no issue since (no warning, no errors, no docker crash etc...). You were definitly right.