Hello,

I'm a user of the 5.0.6 version Plus (last stable version).

In my computer I have 5 hard drives and the usb flash (with unraid on it).

In thus 5 drives, there are :

- 2 samsung of 1 TO

- 3 WD20EARS of 2 TO

You can see my configuration here :

http://hpics.li/b266759

 

The 4 first drives in the list (whose the parity drive) work well.

My problem is when I add the last hard drive for clearing. The clearing begins et reaches 100 %.

But just afterthat, I obtain the configuration which you can see on the link above.

And I can't re-add the disk as disk4 to rebegin cleaning, if I don't reboot unraid.

 

And every time I try, I have this problem.

 

I verified the disk with the tools "Disks" on linux mint (extended test), and I didn't find any problem sector.

 

I appreciate any help.

 

(I have the logs file but it weights 22MO)

 

herbea

(I have the logs file but it weights 22MO)

Try zipping it. Logs compress incredibly well because of all the repeated text.

Yes it works well with this method The file weights 192 + 68 KO.

 

The second part of the zip is below.

logs.zip.001.zip

Please remove *.zip at the end of the files names to read them together.

logs.zip.002.zip

The zip files are broken.

It will be simplier like that.

 

https://www.wetransfer.com/downloads/53e566be377d4c144b7390fa2ee3a46320150228224351/5411caa24eceb79ad1f4f0d5d6e8476420150228224351/532ea8

You definitely have a problem with WDC_WD20EARS-00MVWB0_WD-WCAZA2062430.  The majority of the log is filled up with errors from it.  (This is the drive that was clearing)

 

post up the output from

 

smartctl -A /dev/sdb

 

Are all the drives running off of the motherboard?  What is your power supply?

 

Try reseating the sata cables and the power cables to the drives.

I didn't undertstand where I should enter the command, and what is it interests.

 

1) Are all the drives running off of the motherboard?

 

All drives are directly connected on the sata of motherboard.

 

2) What is your power supply?

 

I have a 400 W Be Quiet 80Plus Gold.

 

3) Try reseating the sata cables and the power cables to the drives.

 

Do you advise me tu unplug all the cables from the drives to verify if the cables work poorly ?

 

Thanks for your help

 

 

I notice that your power supply is not a single-rail variant (as is recommended for use with unRAID), and is comparatively low power.  This may mean that when you add the 5th drive there is not enough spare power available on the rail to the disk drives to power it correctly.  Insufficient power is one of the things that can cause drives to drop off-line unexpectedly.

 

I would be interested to see if others also think this might be a possible cause of your problem?

This evening I will verify the links from all hard drives et how they are allocated between the alimentations.

I notice that your power supply is not a single-rail variant (as is recommended for use with unRAID), and is comparatively low power.  This may mean that when you add the 5th drive there is not enough spare power available on the rail to the disk drives to power it correctly.  Insufficient power is one of the things that can cause drives to drop off-line unexpectedly.

 

I would be interested to see if others also think this might be a possible cause of your problem?

Quite possible. Especially since there are some older 7200RPM drives in the mix.

If it were me, I'd be uncomfortable running that power supply so close to rated capacity. I like to run larger than necessary PSU's so they don't have to work as hard.

I notice that your power supply is not a single-rail variant (as is recommended for use with unRAID), and is comparatively low power.  This may mean that when you add the 5th drive there is not enough spare power available on the rail to the disk drives to power it correctly.  Insufficient power is one of the things that can cause drives to drop off-line unexpectedly.

 

I would be interested to see if others also think this might be a possible cause of your problem?

Absolutely.  The OP was clearing a drive via unRaid (only drive running), then as soon as he tries to add it, the syslog starts filling up with a million errors about the drive (all drives began running).

 

I asked for smart report just as a matter of course.

You definitely have a problem with WDC_WD20EARS-00MVWB0_WD-WCAZA2062430.  The majority of the log is filled up with errors from it.  (This is the drive that was clearing)

 

post up the output from

 

smartctl -A /dev/sdb

 

Are all the drives running off of the motherboard?  What is your power supply?

 

Try reseating the sata cables and the power cables to the drives.

 

I used this command and all values are OK (below Thresh value).

The field WHEN_FAILED is empty for all values.

 

I will verify the cables.

 

I find the consommation for the 2 types of disk (in activity) :

- 3 * WD20EARS : 3 * 5,9 W

- 2 * samsung : 2 * 12,2 W

 

http://www.tomshardware.fr/articles/Samsung-Spinpoint-F1,2-451-14.html

http://www.hardware.fr/articles/imprimer/835/

 

Which type of power supply sould I use so ? (with single ray)

 

I find the consommation for the 2 types of disk (in activity) :

- 3 * WD20EARS : 3 * 5,9 W

- 2 * samsung : 2 * 12,2 W

 

http://www.tomshardware.fr/articles/Samsung-Spinpoint-F1,2-451-14.html

http://www.hardware.fr/articles/imprimer/835/

 

Which type of power supply sould I use so ? (with single ray)

Its not so much the normal power draw for the drives that's important.  Far more important is the spin up power which is going to be a fair amount higher than the 0.5A that your listing as activity draw.  Without looking it up, I usually bank on ~2A per drive spin up power.  So, 2A x 3 drives = 6A.  The Samsung Drives are probably closer to 3A startup.  So you're already up to 12A.  Your supply in theory should be able to handle that, as it should be able to supply 18A on the rail dedicated to the drives.  But, if  you are running case fans, etc also through the molex connectors, then that's going to dwindle.  But it is awfully close to borderline.

 

The net result is that as soon as you finished clearing the drive and then added it in (and all the drives spun up), that's when all the syslog errors happened - which definitely implies power supply.

 

You should be fine with a 450 Watt single rail power supply, depending if you're planning on adding more drives or not.  Single rail because then the power that's currently allocated to the cpu / mobo and not being used the hard drives can take advantage of.  Myself however, I wouldn't go with anything less than 500

Would this power supply be comfortable for my NAS ?

http://www.ldlc.com/fiche/PB00150493.html

FSP AURUM S500 500W 80PLUS Gold

 

I verified the cables et I saw nothing bad.

 

I linked the WD drive which didn't work on the same cable that the two others WD.

The two samsung drives are on other cable.

I will see tomorow morning the result of new cleaning...

Would this power supply be comfortable for my NAS ?

http://www.ldlc.com/fiche/PB00150493.html

FSP AURUM S500 500W 80PLUS Gold

Looks ok for what you have now. Are you planning on adding more disks later?

I don't think taking more than what can support a "Plus" licence.

But why not a sixth drive and a cache disk.

 

The yesterday evening/night gives me a new log file, as if unraid had rebooted after the new test cleaning.

(file log below)

 

What are you opinion ?

logs_unraid_02-03-15.txt

Hello,

I changed my power supply by a more higher and I have always the same problem with one of my hards drives.

My new power supply is below :

http://www.materiel.net/alimentation-pc/enermax-lepa-b650-mb-modulaire-650w-96337.html

 

Would you have any idea ?

If its always the same drive with the issue then the problem is likely the drive.

 

I just rma'ed a wd 3tb that was continually giving me intermittent issues but would always pass any smart test with flying colors