Need Help: Sudden disk detection issue once Western Digital Red disks
Posted 17 November 2012 - 05:31 PM
For over two years I'm running a server with the following configuration with great reliability:
Intel Core i3 530 1.2GHz
4GB DDR3 1333MHz
Intel X25-V G2 40GB (System Drive)
Adaptec RAID Controller 5405, with latest firmware, drivers and management software
Chenbro CK12804 SAS Expander
Norco RPC 4220 case with mini-SAS-to-SATA backplanes
Windows Server 2008 R2
However, this week I installed ten Western Digital Red disks and ever since then I have major problems getting all of my disks to work properly.
My storage configuration:
Adaptec 5405 (Power Saving Options Disabled)
--> Chenbro SAS Expander
--> RAID 6 Array with 10 Samsung F3 EcoGreen SATA II 2TB Disks
--> RAID 6 Array with 10 Western Digital Red SATA III 3TB Disks
After installing the new Red disks, I started building the second RAID 6 array with the same settings used for the Samsung disks. That process went smoothly and the RAID was building slowly. But after only 3 hours, the Controller Alarm went on and all 20 disks were suddenly missing.
After a shutdown and power-on, the controller detected all disks again and continued the building process in addition to verifying the Samsung array for errors. All disks were shown as "optimal" without any hardware related issues. This phase continued without issues for about two days and then the same problem arise and all disks were missing again and I had to start over the building/verifying process again. Since then, this happened twice more.
After that I decided to only verify the Samsung array and stopped all tasks for the WDC array and that worked nicely. So it appears that as soon as the WDC disks are being used, at some point in time, the whole SAS bus will go into an error state and all disk will disappear.
I would really appreciate if you could provide me with some pointers on how to continue trouble-shooting this problem.
Are there any known compatibility issues between the Adapted 5405, CK12804 SAS Expander, RPC 4220 backplanes and/or Western Digital Red disks?
Do I need to use certain settings for the Red disks to work, for example parking timeouts?
Thank you very much for your help.
Posted 22 November 2012 - 05:01 AM
Laptop: Dell inspiron E7440, 8GB RAM, 480GB Crucial M500 mSATA, Win7 Pro
Workstation: Homebrew i3-4350, 16GB, 256GB Micron C400, 3TB Barracuda XT, Win8.1 Pro
NAS: Synology DSM5 running on Pentium G3430, 8GB, 2x1TB HGST & 2x4TB HGST
HTPC:TranquilPC ABEL H2, i3-4010U, 4GB, 64GB M4 mSATA, Win7 Pro
HTPC 2: Pentium G620 on DH61AG in Antec Euler, 4GB, 120GB Kingston V+200, Win7 Pro
Posted 22 November 2012 - 05:54 AM
Can you test the +12 and +5 outputs?
What I have seen on an older system was exactly voltage drop which invalidates all drives in a soft array and the system freezes. Then after a power cycle all seems well until the next big load shows up.
Other than that, I'd give WD a ring on their special NAS support line for the WD Red.
Please note that WD Red are POS (piece of ... ) when they are in large arrays. Even SR has a bench of 4 of them if memory serves me right...
Posted 26 November 2012 - 11:47 AM
I'm also "hoping" for a power issue and I'm currently running a test with only the 10 WDC disks. Haven't had time until today to continue testing. I'm really curious to see if this build-process is able to finish.
I'll report back.
However, if it's a power issue, I find it strange that it just looses all disks but the rest of the system remains usable. It doesn't crash or freeze...
I've contacted all companies involved but so far it's only the usual "check the compatibility charts" advice, nothing useful.
And the WDC Support is as horrible as always in Europe. The 24/7 phone support is not available here and the 24/7 e-mail support redirects me to the normal support e-mail page.
I really hate that Samsung went out of the HDD business
Posted 02 December 2012 - 01:31 PM
With only the 10 WDC drives connected, the controller was able to build the array.
I then re-wired the power cables and so far so good. With all 20 disks up and running, everything seems to be stable now. I hope it stays that way!
Thanks again for your help
0 user(s) are reading this topic
0 members, 0 guests, 0 anonymous users