Dockers Stopped. Unable to write to Docker Image Error - Diagnostics Attached

[JonathanM]

Research the actual tech specs of the RAM, find the certified timing and voltages, then manually set them in the motherboard BIOS.

Test. If errors, bump up voltage by 1 notch, test again. Don't go more than a few percent above stock voltage however.

 

Given the randomness and consistency (yeah, I know) of your errors, I tend to agree that the RAM may not actually be faulty, but stranger things have happened.

[Canes]

OK so here is the latest - just got home.

 

I didn't swap any more sticks around but instead looked up the tech specs for my RAM kit which I found here and manually inputted speeds of DDR4-2133 and 1.2v based on SPD specs. Started memtest with DIMMs 2 & 4 which got the 1 error after 5.5 hours overnight - albeit before manually selecting speed and voltage.  If these 2 sticks at ddr4-2133/1.2v go 12+ hours without errors I assume I need another 12 hours run no errors with 4 sticks in there after that to confirm it was a voltage/speed setting causing RAM issues?  At that point I am OK to start drive replacement process?

 

Hate not having Plex since we cancelled cable, Netflix and everything and only got Prime Video left since that comes free with Amazon Prime shipping.  So spoiled with Plex....

[JonathanM]

18 minutes ago, Canes said:

I need another 12 hours run no errors with 4 sticks in there after that to confirm it was a voltage/speed setting causing RAM issues?  At that point I am OK to start drive replacement process?

At this point, only you know what it's going to take to trust the server again.

 

When you can confidently say, "I KNOW the server can complete a memtest of indefinite duration without error." that's when it's ok to proceed. If you feel it's still sketchy, it's too soon. Think of running the server normally as an extended memtest, only it has real consequences like lost or corrupt data instead of just red lines on the screen.

[Canes]

Just got error after one hour with DIMMs 2 & 4 using SPD settings in BIOS so no dice. Upped voltage to 1.21250 (next available increment in ASUS BIOS) and started another Memtest.

 

If this fails I’ll up voltage another increment and try again but after that I’m going to guess the RAM may not be problem and it looks more like motherboard? I mean what are the chances of 4 sticks all throwing errors in different combinations of 2 and them all starting at once to boot?

[Canes]

Failed with increased voltage again.  Guess I’m ordering new RAM to test with before I go through hassle of new motherboard. Yay…

[Canes]

I ordered new RAM to test with. Linked here.  They are supposed to arrive today so I can start memtest tonight.

 

Ordered a 4 stick kit so I can test up to all 4 DIMM slots as relates to motherboard being a factor using 2 vs 4 sticks.  Will start using 2 sticks and if goes well try all 4 sticks.

[Canes]

New sticks arrived. First two have been installed in slots 1 & 3. and Memtest has been started. Confirmed bios is set at the SPD settings (timing, frequency and voltage) for this memory as well.

 

CPU running cool as cucumber at only 27 degrees haha. They do match black vibe better than the G Skill and always annoyed me on the G Skill one of the sticks always was out of sync with the other 3. Do wish HBA cards were black.

 

About 10 minutes in so far and no errors. Next update will be if this test fails without lasting 12 hours or that it passed 12+ hours with no errors and I am starting next test with 4 sticks.

 

[JonathanM]

1 hour ago, Canes said:

it passed 12+ hours with no errors

Appendages overlapped in your general direction.

[Canes]

2 sticks from the Corsair kit I grabbed on Amazon passed memtest after 14 hours with no errors. Wahoo!

 

Test starting now for all 4 sticks of the new RAM.  Really didn't expect the RAM to pass given challenges with the G.Skill sticks I was using before and all 4 sticks seemingly having issues.  Thought test was going to throw errors and I was going to need a new mobo.

 

Will find out tonight if I'm going to be able to stick replacement disk5 back in tonight once the memtest with all 4 DIMM slots filled completes.

 

 

UPDATE: 3.5 hours into test with all 4 sticks and no errors so far. 

 

UPDATE 2: 6.5 hours into test with all 4 sticks and no errors so far.

Edited May 27, 2022 by Canes

[Canes]

Yay! Memtest ran for over 17 hours and found no errors with all 4 sticks installed.  With that result I went ahead and booted up UNRAID and installed the new drive for disk5. Rebuild process underway - estimated time to complete is 1 day 13 hours.

 

 

I also cleaned up the wiring job in the case and added labels to the hard drives with their respective serial numbers so don't ever have to guess at what disk is in what slot. Before and after pictures attached.

 

Edited May 28, 2022 by Canes

[Canes]

Rebuild completed with 0 errors!

 

Once that completed I went in and disabled dockers via settings (all were stopped during rebuild already), deleted vdisk docker.img and reenabled dockers and then reinstalled the dockers I actually use from previous apps list. Everything appears to be working as it was before it went down now for the past couple hours. Sonarr started grabbing episodes that aired while I was down and they are being moved to the correct Plex folders. Plex is working good as well too - fast playback and remote users are able to play files too.

 

To be safe I downloaded a new diagnostics file.  Anyone seeing anything I should be concerned about in diagnostics?

 

If nothing jumps out in diagnostics file I would say this is problem solved. Thank you so much to @Squid @JorgeB @trurl @itimpi @JonathanM for all your collective help getting me back up and running to this point!!

 

 

jeffunraid-diagnostics-20220529-1801.zip

[JorgeB]

8 hours ago, Canes said:

If nothing jumps out in diagnostics file

Looks clean so far.