charker

Hi

I wonder if anyone has any clues to this problem.

I built a NAS, Server 2008 R2 beta, Adaptec 51645, 8 Hitachi 1TB (.2 revision) drives.

It has worked fine for 2 months. I have it doing the 1/2 spin down after 5 minutes and disks off after 20, if not in use.

Now I have an issue. Randomly the odd disk will not spin-up if they have gone totally to sleep, ie after 20 minutes or whatever.

This breaks the RAID, alarm goes off. I reset the machine go into the Adaptec BIOS config tool, see one drive has gone "gray", it starts to rebuild (onto this "bad" disk) and after 4 hours we all all back OK. Until it happens again.

The first two times it was the same disk, so I put in a new one. A day later a different disk.

I told it to start 1 disk at a time (default is 4) so I could see the culprit. This time CN=1 Dev=0, It starts clicking, alarm goes off...I wait. Other disks on the same CN start up. It sits there "degraded" looking very dodgy to me. I turn off the power and go into the BIOS again. All disks OK!

I reboot, RAID says "impacted". I need to run verify and fix. Another 4 hours!

Between the first failure after fitting the new disk I upgraded Adaptec BIOS and installed latest (matching) 64bit driver and Windows Config software.

It is stable if I leave them at full speed or 1/2 speed but if the go to sleep I'm in trouble.

It could be Windows 2008 R2, it's beta after all but I don't believe that.

Could it be heat related? It is summer (or what passes for that in the UK) after all. I have heat sensors on two of the drives, they average about 38. I also have a fan directed at the RAID card (it used to get damned hot) and a temp sensor near the heat sink now says 40 degrees.

Any ideas appreciated.

raptor_pa

I would think heat or power, 8 1TB drives are going to pull some spinup current, even 4 could. What are the specs of the Power supply?

charker

It's a beefy Seasonic 750W. Even on full power the machine only consumes 180W. The CPU is an AMD Phenom IIe (65W).

I set it to start 1 disk at a time and it still failed.

The bizarre thing about heat being an issue is that these drives fail after they have been "asleep", so no heat in them at all really. Saying that it has been fine for 2 months and when the weather gets warmer it has started to fail!

I have an XCube fitted with 6 temp sensors controlling all the fans. One sensor in each disk stack average between (35 and 40 in warmer weather). The other 2 on the Northbridge (39-44) and the Adaptec heatsink (38-45).

Have read about doing a BIOS 0 upgrade, You disconnect drives, short a pin, upgrade bios and reconnect drives in same order. Could be worth a go?

thanks

Cliff

raptor_pa

The hitachi safe mode jumper, I know it can be done with a PC3000 or Salvation Data Disk Doctor hardware/software tools. I've done it once withe SD's DD to replace a bad mocule in a drive that was failing and just did it for the practice. Not sure if it will work without the hardware part of these tools. I'm sure someone has figured out how to do it, but unless Hitachi has a method I'm always a little dubious about stuff I read on the internet ( he says as he types this into an internet forum :) ) Your mileage may vary, it does sound like it could be a firmware issue, - does Hitachi have updated firmware for your drive? ( Also be aware, certain modules in the firmware can not be rewritten since they contain information specific to the layout of each individual drive so you may get the base upgrade and still not solve your problem )