I have a small array with (4) 3 TB drives.  

 

Last night I copied a few new movies to my "Movie" folder and a few seasons of shows to my "Television" folder.  During this all was well, and as they completed on the array through network explorer I changed the names like I always do to appease PLEX.  At a point when I was about 90% done I noticed that the last two copies (one movie, one show) had error'd out during the copy.  

 

I navigated to the tower GUI and it seemed to be operating very slowly.  I immediately noticed the Docker tab wasn't there and set out to get it back.  I spun down the array and took it offline, and rebooted.  On reboot the tab was back but had no containers or options to add any.  I was sort of focused on this when the timer initiated my mover.  At that point the 256GB SSD had 96.7GB to move.  The interface seemed to actually perk up at this point to the speed I normally expect but that's when it started to go downhill.  

 

I hadn't sorted out the docker issue but I noticed on refresh that the Cache value wasn't decreasing.  I saw a ton of reads on the cache and reads on the parity and array drives.  No writes were being reported.  When I navigated to the Tower through windows explorer I saw all of my shares still.  My pictures, documents, applications folders were all there and contained what they should.  The television and movie folders though were blank.  I checked the array details and the sizes seem like the data should be there.  It was getting late but right before I left it the GUI stopped responding, so did the Tower through Windows explorer.

 

This morning the GUI is still unresponsive but the Tower shows in explorer again, and as before all the folders work but the two.  Now however they don't just show as blank, when I click them in windows explorer it spends a great deal of time thinking (scanning) but still no folders.  Because I can't get in the GUI I can't see whats happening, and the server itself does show drive activity on the LED's, but not the normal write sounds I hear when the platters are working.  Its basically silent but active so I assume its still stuck in reads.

 

I don't want to shut it down before I get an idea of what to expect.  Its been running for about 8 hours now, and I don't mind letting it get through whatever this is as it seems like things are happening.  I don't care if I lose what was cached rebooting but I don't know how long to leave it.  A parity check on my drives (I have about 3.3TB used, drives are HGST 7200rpm) takes about 6.5 hours so its not like this is way out of line.

 

Thanks for any tips,

Ryan

Grab a Diagnostics (Tools->Diagnostics) while the server is having problems and post it here.

Am I safe to reboot the server at the server itself without hurting anything?  Its still showing disk activity (no platter searching though) and is unresponsive to the webgui, but all my shared folders except the TV and Movie folders are accessible via explorer.

Well I gave up and rebooted.  The GUI is up again but very slow.  I see my drive config in the same way.  It shows the correct amount of used space, but the two shares that were to get the files (Movies and TV) are empty in windows explorer.  Also the Docker tab is there but empty.  I upgraded to 6.3.3 while I was in there, no change.

 

I've attached the result of diagnostics.  Very worried as the two shares are orange triangles and within the GUI show nothing in the index.

tower-diagnostics-20170513-1552.zip

Its funny while playing around I decided to disable the Docker and the whole GUI is super fast again.

Syslog is full of ATA errors on this SSD:

 

May 13 15:41:14 Tower kernel: ata2.00: ATA-9: PLEXTOR PX-256S2C, P02637132951, 1.01, max UDMA/133

Replace both cables.

Will give that a try.  I'm glad to have a resource like this with people who know what they are doing!  I'll check back!

Tried a new cable (used but older) same results.  When you say both do you mean power as well?

 

This is my first SSD in over 5 years when I had my first one fail within 6 months and I had hoped things had come a long way.  Sure hope I don't have another dud.

Yes, power as well, if still errors try a different SATA port, if all else fails it's probably a bad SSD.

Well I swapped power and cables with one of the HDD's and from my limited ability to translate the error logs it looks to me like the errors stayed with SDC which is the SSD Cache.  

 

So what do I do now?  My two shares on the main array are still empty when they contained about 2TB of stuff yesterday, despite the fact that the GUI reports the space is still gone.  Can I run the array without the cache drive to see?

 

This is the new error log.

tower-diagnostics-20170513-1709.zip

5 minutes ago, ryankenn said:

Can I run the array without the cache drive to see?

 

Yes, errors persist, try that

Swtiched the SSD to another port.

tower-diagnostics-20170513-1723.zip

Removing the cache drive returns all my stuff which is great.  Mad about another dead SSD though.

 

I also still have no Docker.  The tab is there and while it shows that it is "ON", its status is STOPPED so no Plex.

You'll need to re-create the docker image, if you have a backup of appdata restore it so you don't need to reconfigure them all.

 

If you use VMs you'll need to do the same with libvirt.img

Nice thanks.  I deleted the image and it came back.  Don't know if I have a backup of appdata though, I'll have to check.

 

Thanks for you help, its very much appreciated.  Now I get to fight with NCIX to see if they can RMA a 5 month old SSD>