Hi folks,

today i changed my last 2TB-Disk - a WD-Red - against a new 4TB WD-Red. At about 40% of reconstruction, i saw a significant slowdown.

I made a Screenshot and like to ask, if someone saw this behavior on his sys as well.

Thanks

Edited May 17, 20179 yr by Zonediver

Yes, more than once, looks like a WD getting slow zones.

4 minutes ago, johnnie.black said:

Yes, more than once, looks like a WD getting slow zones.

 

Ah ok. This explains why i saw this during my last reconstruct of a new WD-Red 4TB two months ago.

Is there a chance to find out, which disk has this "slow zones" on it?

Thanks jonnie for your reply

 

Edited May 17, 20179 yr by Zonediver

You can usually find which disk it is with the diskspeed script, increasing the number of test points helps.

2 minutes ago, johnnie.black said:

You can usually find which disk it is with the diskspeed script, increasing the number of test points helps.

I did this before i changed the disk but there was nothing unusual. I will test this again with more "Testpoints".

Edited May 17, 20179 yr by Zonediver

I usually use 101 test points for this purpose, it tests each disk every 2%, my last one with this problem:

Yea that looks terrible...

My reconstruction is at 49% and still slow - WTF

The question is, if i should change the disk or not...

 

EDIT: Now it passed 50% and the rebuild-process is fast again.

I must test it to see whats wrong.

Edited May 17, 20179 yr by Zonediver

So, short reply from me:

I did a test of the affected disk and found this.

Shall i be scared about or is this a minor problem?

The SMART-selftest says that the Disk is ok and has no errors - BUT...

The Raw read error rate is increasing slowly 1-2 per week

I allready changed the SATA-cable but the raw read error goes still up.

I would be thankful for any advice.

 

Edited May 18, 20179 yr by Zonediver

Those normalized SMART attributes are still at max 200, so they should be of no concern, still was looking at my last 2 disks with these issues and both also show errors there while another 8 WD disks I can check ATM show 0 RAW for both, let me check my big server and see if all good disks are at 0, it's possible these SMART RAW values are a clue.

 

As for replacing it, when I find a disk like this, I usually replace it when possible, since IMO they will get worse until they eventually fail and performance, specially if using turbo write, will be noticeably affected.

Ok, i see. Then i will keep an eye on this WD-Red and see what happens after a while.

Thanks for your help johnnie.black - good work

Look at this...

I think, this Disk is crap now

Read-Speed at the end: 0.1MB/s...

 

Edited May 23, 20179 yr by Zonediver

Yes, that's why I usually replace a disk with slow sectors ASAP, in my experience they get worse quickly, and some of those slow sectors turn to bad sectors.

Oh, and as for the Raw Read Error Rate most of my WDs are at 0, but some are not and there are no signs of any issues for now.

Thanks for your reply johnnie - all my WD Red's have 0 RRE except this one... never mind, the disk is two years old - i change it.

Edited May 23, 20179 yr by Zonediver

Is there a way to do tests like this while it's in the array?

Sent from my Nexus 6P using Tapatalk

23 minutes ago, pinion said:

Is there a way to do tests like this while it's in the array?

Sent from my Nexus 6P using Tapatalk
 

 

Diskspeed plugin can test all array (and unassigned) devices:

 

 

Thanks! I had found the script. Is there a proper plugin I missed?

Sent from my Nexus 6P using Tapatalk

36 minutes ago, pinion said:

Thanks! I had found the script. Is there a proper plugin I missed?

Sent from my Nexus 6P using Tapatalk
 

 

Yes, my bad, I was going to link the plugin:

 

 

Thanks! I forgot all about being able to search in the apps tab... Server has been off for awhile