My Readynas Duo V2 of a couple of months age was initially loade with two Hitachi 500GB drives, one Deskstar and one Videostar (my favourite). These filled up and I bought two Seagate ST1000DL002-9TT153 1TB drives. I loaded first one, synced, then the other, synced again. No problems. With both drives in sync and running at a comfortable 38 C I reckoned all was well.
Maybe a week later the first (LH) drive dropped out. Log said it was removed. Pull the drive and re-insert, drive recognised and resyncing. Resync completes without error. No errors logged on drive. Maybe a month, and it has just dropped that drive again. Pull it, re-insert, recognised, re-syncing. No errors logged on drive.
Now is it the drives or the NAS?
I will comment that I have over the years collected piles of bricked drives and the proportion and Seagate and WD put me off buying those brands for a long time. It has only been prices which swayed me this time, and owner changes at Hitachi.
It looks like the ST1000DL002 is not on the HCL for the Duo V2. That leaves me suspecting the drive. Try pulling it out and running the manufacturers tools on it to see if it is giving any errors. Additionally I would suggest backing up all of your data as soon as possible as the resolution may be changing out the drives and one drive failing brings the other one into question.
Yes, I have studied the HCL, and note that the same problems afflict drives on that list. The list of drives is also rather limited and specific type numbers are a pain to source at reasonable costs due to manufacturers inconveniently changing their product ranges frequently.
It used to be recommended that drives in a RAID array be of mixed heritage to lessen the possibility of more than one drive exhibiting the same fault at around the same time. This is not a Netgear recommendation either.
Gone are the days when a new drive meant patching and burning a new BIOS chip. I would expect drives conforming to a set standard to play well with machines designed to that standard. It may be wise to keep to the same revision of that standard when mixing disks, say ATA8.
I chose the Readynas Duo V2 for a number of reasons. The CPU was fast enough to handle a couple of disks and a fast network. My own experiments used Celeron 1GHz; 1.3Ghz; 1.4GHz; 2.4GHz. I wanted a closed box I could not fiddle to death. My usage is low. It gets switched on maybe once or twice a week and a few GB's of data will be copied for safe storage.
When money allows the Seagate disks will be dumped. I doubt if the replacements will be on the HCL either. They may even be of different manufacturers.
Extended offline - completed without error. Conveyance offline - completed without error. Short offline - completed without error.
This is what I suspect is happening....
Man goes to hospital. Doc reckons he needs a good feed. Man stuffs himself silly and decides to have a doze while he digests his meal. Man wakes up in a mortuary!
Summary, the NAS driver times out before the drive buffer signals it is ready for more data. The driver then assumes the drive is dead.
Installed ssh which allowed me to check what the Readynas systems made of upgrading from a 500GB mirror to a 1TB mirrored array.
#################################################### root@readynas:~# mdadm --detail /dev/md0 /dev/md0: Version : 1.2 Creation Time : Sat May 12 10:23:31 2012 Raid Level : raid1 Array Size : 4193268 (4.00 GiB 4.29 GB) Used Dev Size : 4193268 (4.00 GiB 4.29 GB) Raid Devices : 2 Total Devices : 2 Persistence : Superblock is persistent
Update Time : Mon Jul 2 13:41:22 2012 State : clean Active Devices : 2 Working Devices : 2 Failed Devices : 0 Spare Devices : 0
Name : E0469AA09AE7:0 UUID : 15b6b5d0:b3940c96:f7896447:ac4a0bb8 Events : 257661
Number Major Minor RaidDevice State 3 8 1 0 active sync /dev/sda1 2 8 17 1 active sync /dev/sdb1 root@readynas:~# cat /sys/block/sda/queue/logical_block_size 512 root@readynas:~# cat /sys/block/sdb/queue/logical_block_size 512 root@readynas:~# cat /sys/block/sda/queue/physical_block_size 512 root@readynas:~# cat /sys/block/sdb/queue/physical_block_size 512 root@readynas:~# ################################
Whether the fact that it thinks it has a 4TB array and not a physical 1TB array is causing problems, or whether the Seagate internal translation of 4k block size to 512 byte block size is causing hiccups I do not know. Any wise men out there? What should it say?
PS I meant Cinemastar not Videostar. Blame Buggles. They run up to 5C cooler than Deskstar at the same speed, and I had no problems using them in a FreeNAS mirror.
Whoops!!! you are right. I neglected the other three md devices, and fitted what I expected to see onto what was reported. Still a little miffed at having to get to grips with the internal organisation of what I bought as an appliance. The other bulkier and less environmentally sound option was a PC based NAS using Slackware with a 2.4GHz Celeron. At least then I would understand exactly what is in there and what it is doing.
