extremely slow parity check. from around 10.5hrs to a reported 2days+

[mctavish01]

Ok, gonna spew out a bunch because it sure was a crazy morning. 

Last night I had just noticed that the server was acting up and couldn't connect to it at all so I had to force shutdown on case. When I turned it back on the first time, it did not come back online. so had to force shut down again. Took a 3rd try for it to come back online. I figured something was up but didn't have time to do any trouble shooting as it was already late so I just shutdown from web interface. 

This morning when I went to boot it up, it didn't. tried a few more times to no success so I grabbed the HDMI to hook up and actually see what was going on.
It kept getting stuck in boot at "unpacking initramfs", searched on here and google while I did a memtest. The test did not finish and it just randomly rebooted on it's own and went back to "unpacking initramfs". Shut the system down and took out a stick of RAM and it did manage to boot.
Now all this above may have nothing to do with the slow parity check, but it has all happened in the same day so figured it was worth mentioning.

once everything was up and running with 8gb of ram down from 16gb it started a parity check, which in itself I thought was strange as it is on a schedual for end of month but figured it's maybe noticed a change and just checking for safety.

I have x8 Seagate ironwolf 6tb drives with a single parity which normally only takes around 10.5hrs, this time it is saying it will take 2days + and is going extremely slow. 
Now none of the drive have any warnings going off anywhere. What would be the best course of action to figuring out what is going on?


Due to the earlier RAM issue I have decided to source out some newer hardware that has been on the back burner for a while. however I would love to know the integrity and performance of the drives and system are still in good order before I make any more changes.

Thank you for any and all help, response may be slow as heading to work.

TLDR: Possible unrelated RAM issues in the morning, but during parity check my normal time for my array has gone from 10 hours to 48+

[Frank1940]

Get a Diagnostic file      Tools    >>>   Diagnostics      and upload it.   Shutdown any Dockers and anything else that is running until you figure out what is going on.  Try not to do any writes to the array that is not involved with the Parity check.  

 

I suspect that you have a hardware problem.  It could be memory, PS, MB or CPU.  

[mctavish01]

unraid-diagnostics-20190711-1707.zip

See attached.

[mctavish01]

1 hour ago, Frank1940 said:

Get a Diagnostic file      Tools    >>>   Diagnostics      and upload it.   Shutdown any Dockers and anything else that is running until you figure out what is going on.  Try not to do any writes to the array that is not involved with the Parity check.  

 

I suspect that you have a hardware problem.  It could be memory, PS, MB or CPU.  

also the server is powered down until i know whats up. also have new ram/cpu/mobo on the way already

[Frank1940]

When you build the new system, I would not connect any drives to it until I had run a memtst for 24 hours first.  That would help eliminate the possibility that the PS is causing problems.  

 

I had a quick look at the Diagnostics file and noticed that you appear to have started up the server only long enough to capture the diagnostics.  As a result I didn't see any problem as it never had time to act up. 

 

EDIT:  You are probably wise to simply replace the CPU, MB, and RAM as finding these types of problems can be difficult...

Edited July 12, 2019 by Frank1940

[mctavish01]

6 hours ago, Frank1940 said:

When you build the new system, I would not connect any drives to it until I had run a memtst for 24 hours first.  That would help eliminate the possibility that the PS is causing problems.  

 

I had a quick look at the Diagnostics file and noticed that you appear to have started up the server only long enough to capture the diagnostics.  As a result I didn't see any problem as it never had time to act up. 

 

EDIT:  You are probably wise to simply replace the CPU, MB, and RAM as finding these types of problems can be difficult...

would it be advantageous to start up and start a parity check then grab a new diag to see if anything shows up?
And thanks for the tip on the test cycle I will def do that.
 

[mctavish01]

I feel it's safe to say we found the culprit? 

unraid-diagnostics-20190712-0421.zip

[Frank1940]

Open up the case.  Clean it out with compressed air.  (Do this is a place when the dirt won't be a problem!!!)  Pay particular attention to the CPU cooling fins.  Check to see that the fans are all running.  The CPU fan would be a prime suspect,.  

[mctavish01]

30 minutes ago, Frank1940 said:

Open up the case.  Clean it out with compressed air.  (Do this is a place when the dirt won't be a problem!!!)  Pay particular attention to the CPU cooling fins.  Check to see that the fans are all running.  The CPU fan would be a prime suspect,.  

it's actually extremely clean in there =\ I blow out all my system twice a year.
Leaning more towards a possible bios issue at this point. That or maybe the pump is not working as it should. it's on older i7-4770k with a 120 Corsair AIO (forget the model) but at this point yea totally worth a clean to rule it in or out. Also wonder if I have a random stock cooler that would fit to test.

[Frank1940]

Are you over-clocking this computer?  It is a general recommendation not to overclock Unraid.  (You want a stable a platform as possible when you are running a server!) 

[mctavish01]

4 hours ago, Frank1940 said:

Are you over-clocking this computer?  It is a general recommendation not to overclock Unraid.  (You want a stable a platform as possible when you are running a server!) 

no the cpu had never been over clocked. It was a old "first" build were you just buy the "best" everything whether you are ever gonna use it's potential or not