Still have good quality older sata hardware raid cards that require 512 bytes/sector. As far as I know HD manufacturers are not making native 512 bytes/sector drives any more.
Some have better 512e emulation than others. Looking for some advice on which to avoid and which are recommended. Thanks. PS this is for a CentOS6 server.
On 2/27/2015 12:53 PM, Robert Arkiletian wrote:
Still have good quality older sata hardware raid cards that require 512 bytes/sector. As far as I know HD manufacturers are not making native 512 bytes/sector drives any more.
Some have better 512e emulation than others. Looking for some advice on which to avoid and which are recommended. Thanks. PS this is for a CentOS6 server.
any of the 'enterprise' nearline storage or NAS drives should be fine. I wouldn't use anything else in a RAID setup.
Seagate NS series, for instance, or WD Red or Re, etc.
On Fri, Feb 27, 2015 at 1:53 PM, Robert Arkiletian robark@gmail.com wrote:
Still have good quality older sata hardware raid cards that require 512 bytes/sector. As far as I know HD manufacturers are not making native 512 bytes/sector drives any more.
512n drives still exist, although they tend to be a bit smaller, 2TB or less. http://www.hgst.com/tech/techlib.nsf/techdocs/FD3F376DC2ECCE68882579D40082C3...
4Kn drives are appearing now also. I don't expect these drives to be bootable except possibly by systems with UEFI firmware. It's also possible hardware RAID will reject them unless explicitly supported. http://www.hgst.com/tech/techlib.nsf/techdocs/29C9312E3B7D10CE88257D41000D8D...
Some have better 512e emulation than others. Looking for some advice on which to avoid and which are recommended. Thanks. PS this is for a CentOS6 server.
The emulation implementations don't come into play if the alignment is correct from the start. The better implementations have significantly less pathological behavior if alignment is wrong, but that's anecdotal, I don't have any empirical data available. But I'd say in any case you want it properly aligned.
Chris Murphy wrote: <snip>
The emulation implementations don't come into play if the alignment is correct from the start. The better implementations have significantly less pathological behavior if alignment is wrong, but that's anecdotal, I don't have any empirical data available. But I'd say in any case you want it properly aligned.
You really, really want it properly aligned. We ran into that problem when we started getting 3TB drives a couple-three years ago. Proper alignment made a measured... trying to remember, but I think it was at *least* 20% difference in throughput.
Alignment's easy: using parted (the user-hostile program), if you do go in with parted -a optimal /dev/drive, and do mkpart pri ext4 0.0GB 100% (for non-root drives, for example), it's aligned correctly.
mark
On 2/27/2015 3:06 PM, m.roth@5-cent.us wrote:
Alignment's easy: using parted (the user-hostile program), if you do go in with parted -a optimal /dev/drive, and do mkpart pri ext4 0.0GB 100% (for non-root drives, for example), it's aligned correctly.
i found -a optimal to do weird things, and almost always complain. I just use -a none now, and specify partition start in (512b) sectors, like..
# parted /dev/sdc align none mklabel gpt mkpart pri 512s -1s
don't start at 0, as thats where the MBR or GPT has to go. 512 sectors is 256K bytes, which puts you on a erase block boundary with most SSD's as well as HD's. -1s is end of the disk.
The default for fdisk, parted, and gdisk is starting the first partition on LBA 2048, which is 8 sector aligned. You don't need any options. The alternative is to simply not partition the drives or the resulting RAID and just format it.
Chris Murphy
On Fri, Feb 27, 2015 at 2:59 PM, Chris Murphy lists@colorremedies.com wrote:
On Fri, Feb 27, 2015 at 1:53 PM, Robert Arkiletian robark@gmail.com wrote:
Still have good quality older sata hardware raid cards that require 512 bytes/sector. As far as I know HD manufacturers are not making native 512 bytes/sector drives any more.
512n drives still exist, although they tend to be a bit smaller, 2TB or less.
http://www.hgst.com/tech/techlib.nsf/techdocs/FD3F376DC2ECCE68882579D40082C3...
I too noticed that HGST (now owned by WD) makes native 512n drives. That pdf states that they come in 2,3,4 TB models. (A6 in the model # represents 512n). But there are almost no reviews on these HGST native 512n drives online.
4Kn drives are appearing now also. I don't expect these drives to be bootable except possibly by systems with UEFI firmware. It's also possible hardware RAID will reject them unless explicitly supported.
http://www.hgst.com/tech/techlib.nsf/techdocs/29C9312E3B7D10CE88257D41000D8D...
Some have better 512e emulation than others. Looking for some advice on which to avoid and which are recommended. Thanks. PS this is for a
CentOS6
server.
The emulation implementations don't come into play if the alignment is correct from the start. The better implementations have significantly less pathological behavior if alignment is wrong, but that's anecdotal, I don't have any empirical data available. But I'd say in any case you want it properly aligned.
According to this pdf [1] alignment is important but from what I understand 512e emulation still has a small RMW performance hit from writes that are smaller than 4k or if the writes are not a multiple of 4k.
Also it's probably not a good idea to mix 512e with 512n in a raid set. Although this may be hard to avoid as drives fail in the future.
[1] http://i.dell.com/sites/doccontent/shared-content/data-sheets/en/Documents/5...
On Sat, Feb 28, 2015 at 12:33 AM, Robert Arkiletian robark@gmail.com wrote:
According to this pdf [1] alignment is important but from what I understand 512e emulation still has a small RMW performance hit from writes that are smaller than 4k or if the writes are not a multiple of 4k.
There shouldn't be writes smaller than 4KB since ext234, XFS, and Btrfs all use 4KB block sizes. There is a possible case where the XFS journal writes are 512 bytes, this can be fixed by specifying a 4KB sector size at mkfs time if it's not auto-detected.
Also it's probably not a good idea to mix 512e with 512n in a raid set.
Scrambled eggs mixed with yogurt? Offhand it doesn't seem like a bad idea (won't kill me), even if it also may not be a good idea (sounds suboptimal).
-
Chris Murphy -
John R Pierce -
m.roth＠5-cent.us -
Robert Arkiletian