Tony wrote:
Hello JDK,
Your server uses an AMD SB820M for SATA support, so that shouldn't be a problem - the ahci driver handles that hardware out of the box.
The drive you are using however is a different story. Couple of things you may need to address if not done so already:
<snip>
Tony -- thanks for the detailed info. I did experiment with using jumpers to limit the drives to SATA-2. I'll check whether they are limited at the moment. And I did tweak the disks for the head parking issue (like many WD users I'm used to doing that as a matter of course now).
Ralph wrote:
As far as misaligned partitions, I'm really not seeing that from the filesystem details, we're using 4K blocks and starting the data partition on block 0
Quote:
I am pretty sure that the partitioning code used by NASLite-2/M2 is sound.
On that note, after looking at the filesystem details from your drives, there's a couple of really odd things.
Default mount options: user_xattr acl
Default directory hash: half_md4
Neither of those options was set by NASLite. ACL's aren't supported at all by the NL kernel.
Ralph -- the disk info I posted above is the data after I re-partitioned the disks with gdisk (i.e. I re-partitioned them outside of NASLite). All I can say is that, as I said in a previous post, transfer speeds jumped markedly after I did so. It may just be a particular hardware combination that is causing issues in my case.
If it's any help, this is what fdisk said about both of my disks prior to being re-partitioned in gdisk:
https://dl.dropboxusercontent.com/u/13895942/fdisk-partition-data.pngAs I said, I'm no expert in these matters, but I have read elsewhere that for Advanced Format disks, the start parameter must be divisible by 4, which it was not for the 3TB disks, according to fdisk.
And this is what gdisk said (about both disks) prior to being re-partitioned (it claimed the GPT was corrupt):
https://dl.dropboxusercontent.com/u/13895942/GDisk-GPT-damaged.pngSorry I did not post these screengrabs before, it slipped my mind.