[SOLVED] Large Number of Writes on Cache Drives

[Limy]

I thought I would throw this out, to see if anyone else is seeing the same thing.

 

Ever since I setup my unRaid server (early this spring) , I have seen a large number of writes on my SSD cache drives.

 

Currently, it has only been running for 2 days, 15 hours and these are the current values:

 

 

Granted, I do have the following on the cache drives:

 

 

I have two VM's and a few dockers running.

 

I also have the trim plugin Dynamic SSD Trim.

 

I'm just wondering if it's normal to see that high number of writes or maybe I have something setup wrong.

 

Thanks for any feedback.

[Squid]

Read/Write counts aren't 100% accurate under any circumstance.  Different controllers, etc will all give different results under identical conditions.

 

But, everything is all relative.  Windows as an example isn't exactly known to not write to hard drives.  Docker applications, would be similar, depending upon the container, what its doing, etc.

[Limy]

2 minutes ago, Squid said:

Read/Write counts aren't 100% accurate under any circumstance.  Different controllers, etc will all give different results under identical conditions.

 

But, everything is all relative.  Windows as an example isn't exactly known to not write to hard drives.  Docker applications, would be similar, depending upon the container, what its doing, etc.

So basically, nothing to be concerned about.

[Squid]

My writes on the cache drive for 1 day, 20 hours is 958,577

 

1 LibreElec VM and a bunch of containers.  You're almost twice as much over twice as long a period.  I don't think its anything to worry about.

 

Kinda like telling your dealer that your odometer has increased by X amount of miles.  Their response would be that you drove it X miles.   Now if your odometer increased by a hundred thousand in a short time period, then there's a problem (or your kids have way too much fun)

[Limy]

Well I'm glad my odometer doesn't go up like my cache otherwise I would need a new vehicle a lot sooner

 

[pwm]

On 6/15/2018 at 7:35 PM, Squid said:

Read/Write counts aren't 100% accurate under any circumstance.  Different controllers, etc will all give different results under identical conditions.

 

But, everything is all relative.  Windows as an example isn't exactly known to not write to hard drives.  Docker applications, would be similar, depending upon the container, what its doing, etc.

 

Isn't the read and write counts picked up from /proc/diskstats which means it's what the OS thinks and not that the disk controller may think.

[strike]

On 6/15/2018 at 7:44 PM, Squid said:

My writes on the cache drive for 1 day, 20 hours is 958,577

 

What about 5.186.249 in 15 hours, Is this something to worry about?   I recently went from two 120 GB cache drives to one 1 TB drive. And installed a windows 10 gaming VM. I never noticed much writes when I run my linux mint vm, but with windows it just seems to sky rocket.. The count literally goes up every second!  I haven't turned off hibernation yet because I'm using it. But I will try to disable it to see if that makes a difference. Something must be wrong, that's for sure..