Well supported, reliable NICs for Redhat Linux/Fedora?

Alan Cox alan at redhat.com
Wed Nov 19 21:23:12 UTC 2003


On Wed, Nov 19, 2003 at 03:51:55PM -0500, Lamar Owen wrote:
> It's not read cache only that is the issue.  It's the idea that when the 
> application tells the kernel 'don't return until the data has hit the disk' 
> that the kernel can tell the drive 'get back to me when the data has hit the 
> disk' and the drive then can notify the kernel of that fact while still 

Thats read cache only + TCQ

> processing other reads and writes without messing up a nice elevator, but 
> keeping the writes in the order given.  I (being the PostgreSQL backend, for 
> instance) must be able to be sure that what I have written to the disk is 
> actually written to the disk in the order I specified.

TCQ is in ATA6 in some SATA drives.

> What is needed is the full FUA extensions  slated for ATA-7.  In the case of 
> PostgreSQL, it is urgent that the WAL gets written before the actual data 
> page on the disk.  See 
> http://www.ussg.iu.edu/hypermail/linux/kernel/0304.1/0450.html for the actual 
> text of what the Maxtor guy had to say.

Yeah. I spent quite a bit of time talking with drive vendors. The drives have
metadata and in essence nowdays its a storage appliance with a file system
and the whole works on it. In fact if you put a modern disk drive in an old
PC its quite possible the PC is the slower CPU.

To get back on topic

2.4 SATA layer over SCSI (eg Promise drivers) do know how to use TCQ and
controller level queueing so can make good use of the drives that have it.
The parallel IDE in 2.4 kernels can't so it may be a 2.6 feature (FC2) 
although PATA TCQ is so mindbogglingly screwball its a very good
candidate for "never"

Alan





More information about the fedora-test-list mailing list