Green Drives and Linux Software RAID

[Max|m]

Has anyone had any experience with green drives on Linux software RAID? Will it work? Is it worthless? Does it go out to lunch randomly?

I read someplace that you don't have to use RAID Edition TLER green drives with Linux software RAID. The discussion also said that at worst the RAID Volume would hang until the disk doing error recovery comes back. I can deal with a little pause every now and then but I want to confirm it works at all before investing a bunch of money into it. Does anyone have any experiences at all with this?

Background
I'm thinking of setting up a home storage server and was hoping I can at least get away with software RAID 5 or 6 on green drives. I know everyone loves WHS right now but I need an ISCSI target server for two VMware ESX servers. The files I determine are super important will be backed up to a encrypted internet backup/storage location. I will just have to hope the rest doesn't go up in a puff of smoke. I would also rather not have to figure out what data is missing and have to get it again after a hard drive failure, hence the use of RAID. I know RAID is just availability and not a backup but I can deal with losing the data, I just don't want to.

General Specs
Openfiler
Norco 4220
2 x 250GB WD2500BEKT - Hardware RAID 1 (OS and ISCSI Share)
10 x 2TB WD20EADS - Software RAID 5 (Movies/Pics/Computer Backups/Etc..)
Will eventually add another 10 disks and convert to RAID 6 when needed

Thanks,
Max|m

 

[lightp2]

Disclaimer : I am not an experience user.

1. You mentioned you can tolerate long delay, experiment with non-TLER drives
2. You mentioned iSCSI and ESX

So one key question is the timeout issue in case the system goes into deep recovery. I found an article that may offer some relevant info.

http://www.tuxyturvy.com/blog/index.php?/archives/53-iSCSI-Boot-from-SAN-with-VMware-ESX.html

I would also like to learn whether Windows-guestVM itself can sustain long iSCSI storage timeout. However, this should not be a major problem for you since this is a home server setup.

Another side question is usually standard motherboard has only 6 SATA ports. More premium one will have 8 ports. To spend huge money for mb with10-12 SATA ports defeat the purpose because those are usually top boards for enthusiast/overclockers and for your purpose you might as well save the money towards buying a good disk controller, which then brings us back to the original question. The cheapest non-raid areca ARC-1300ix-16 is already USD370, since you perhaps need to buy expensive controller card, maybe just pay extra to get the hardware raid function?

Note: myself running on Fedora 12/SWRAID/NFS datastore for ESXi

 

[williamlee]

I run mdadm (Linux raid) with green drives, and it works just fine for me. I have not had any problems using these drives in an array. Let me know if you have any other questions.

 

[don_xvi]

I have 4 GP 1TBs in a software RAID 6. It all works just fine, but I haven't (as far as I know) had any of the drives go into a deep error recovery. Note that at least some GP drives (I think it was only the older EACS models) try to power down so aggressively that they'd spin down, only to spin back up for some sort of (EXT3?) journaling operation so they rack up the head park cycles quickly, and they're rated for a finite number of those. There was a utility that WD introduced to allow one to increase the time before the drives would spin down to prevent this issue and it's worked just fine to keep my oldest two drives spinning happily. Note that the energy burning ulprit here is the journaling activity of the file system.