Date: Fri, 13 Feb 2004 14:48:42 +0000 (UTC)
From: Ramesh K <kram@astron.nl>
To: linux-xfs@oss.sgi.com
Subject: more experiments with realtime subvolumes.
In-Reply-To: <Pine.LNX.4.44.0402131210350.6824-100000@wop10.nfra.nl>
Message-ID: <Pine.LNX.4.44.0402131420170.8002-100000@wop10.nfra.nl>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Sender: linux-xfs-bounce@oss.sgi.com
Errors-to: linux-xfs-bounce@oss.sgi.com
Precedence: bulk


Hello:
	Some more experiements with the realtime subvolumes. Please do pass your
comments. Before we start, my system (OS - SuSE 8.1, kernel 2.4.19-64GB-SMP,
hardware: Transtec, dual-xeon 2.8 GHz, with 3ware Raid Controller - 8*120GB IDE
disks on controller), doesn't seem to have O_DIRECT and O_LARGEFILE options. I'm
downloading a newer kernel (2.4.21) from suse website - hopefully that has these
options enabled.
	My harddisk is setup as explained in my previous mail, 100 GB
meta-data section, and 800 GB realtime section. More questions on realtime
subvolume:
	When I open a file and try to place it in the realtime section with:

struct fsxattr myfsxattr;
struct xfs_flock64 file;

fd = open("/data/kram/data/file1.dat",O_CREAT|O_RDWR);
myfsxattr.fsx_xflags =  XFS_XFLAG_REALTIME;
ioctl(fd,XFS_IOC_FSSETXATTR,&myfsxattr);
file.l_whence = 0;
file.l_start = 0;
file.l_len=2*1024*1024;
ioctl(fd,XFS_IOC_RESVSP,&file);
close(fd);

	And then followed it up with, ioctl(fd,XFS_IOC_FSGETXATTR,&myfsxattr);
This call returns fsx_xflags as 0x00 (meaning no realtime, and not
preallocated). Is it because of the absence of O_DIRECT in the open call?
	What are the alignment constraints? The man page of xfsctl explains that
"d_mem is the memory alignment requirement of the user's data buffer.
d_miniosz specifies block size, minimum I/O request size, and I/O alignment.The
size of all I/O requests must be a multiple of this amount and the value of the
seek pointer at the time of the I/O request must also be an integer multiple of
this amount. d_maxiosz is the maximum I/O request size which can be performed on
the file descriptor." Does this mean a write of "char array[1024*1024]" is
aligned with 512-byte boundary for the I/O to succeed? Or in other words, is
array[1024*1024] d_mem aligned, or do I have to do something special to get it
aligned?
	Thanks for your responses.

-Ramesh

>
> Hi:
> 	Thanks Nathan. I wanted to use ramdisk only to store filenames, and the
> actual data in realtime subvolume. Thoguht it'll have better performance.
> 	However I've now repartioned my disk to have two partitions: one of
> 100GB and the other of 800GB (/dev/sda1 and /dev/sda2 respectively). I'm using
> /dev/sda1 as my regular xfs volume (log plus metadata), and configured /dev/sda2
> as the realtime subvolume. This is the output of mkfs.xfs:
>
> dop93:~ # mkfs -t xfs -f -d unwritten=0 -r rtdev=/dev/sda2 /dev/sda1
> meta-data=/dev/sda1              isize=256    agcount=25, agsize=1048576 blks
>          =                       sectsz=512
> data     =                       bsize=4096   blocks=25601577, imaxpct=25
>          =                       sunit=0      swidth=0 blks, unwritten=0
> naming   =version 2              bsize=4096
> log      =internal log           bsize=4096   blocks=12500, version=1
>          =                       sectsz=512   sunit=0 blks
> realtime =/dev/sda2              extsz=65536  blocks=208836967,
> rtextents=13052310
>
> dop93:~ # mount -t xfs -o rtdev=/dev/sda2 /dev/sda1 /mnt
>
> 	It's been good so far. Now how do I determine the amount of free space
> in my realtime subvolume? Do I need to have the realtime subvolume on another
> device altogether (instead of another partition) to get better performance?
> 	I don't know what version of XFS I'm using (it's not v1.3, but the one
> packaged with SuSE 8.1). I have new version ( 2.2.1-28) of xfsprogs though.
> 	In the mean time I also tried to write a 'realtime' file on the realtime
> subvolume - but SuSE 8.1 seems to have disabled O_DIRECT and O_LARGEFILE
> support. Any clues on that? I'm trying hard to get the realtime subvolumes
> working - hopefully it is more deterministic than the regular buffered I/O. Any
> info on these lines would be great. Thanks in advance.
>
> -Ramesh
>
> > On Thu, Feb 12, 2004 at 03:55:10PM +0000, Ramesh K wrote:
> > >
> > > Hello,
> > > 	I'm trying to use realtime subvolumes under Linux (SuSE 8.1, running
> > > kernel 2.4.19-64GB-SMP (I know it is very old!!)
> >
> > You should be aware that realtime support is marked experimental
> > and may not function correctly.  That said, Eric made several
> > fixes in this area awhile ago and the last time I tested it, it
> > was looking much more stable.
> >
> > > 	The hardware: I'm using a 3ware RAID controller, and am using RAID level
> > > 0 (all eight 120GByte disks together are seen as one disk of 960 GByte device).
> > > 	Now I want to use this entire 960 GBytes for my realtime data, and want
> > > to place metadata on ramdisk. When I try:
> > > dop93:> mkfs -t xfs -f  -d unwritten=0 -r rtdev=/dev/sda1 /dev/ram1,
> > >
> > > 	I get warnings on block size of ramdisk. And when I try to mount the
> >
> > What is the warning?
> >
> > > device with: mount -t xfs -o rtdev /dev/sda1 - the call never returns.
> > > 	Am I doing something very gross? Thanks.
> >
> > That mount line is incorrect, you need to use something like:
> > 	mount -t xfs -o rtdev=/dev/sda1 /dev/ram1 /mnt
> >
> > It seems odd to be using a ram disk in this way...
> >
> > cheers.
> >
> > --
> > Nathan
> >
>
> **************************************************************************
> Ramesh K
> (Mo/Tu): Radiosterrenwacht Westerbork, Schattenburg 1, 9433 TA Zwiggelte.
> ph: +31-(0)593-598755 fax:+31-(0)593-592486
> (We/Th/Fr): Stitching Astron, Oude Hoogeveensedijk 4, 7991 PD Dwingeloo.
> ph: +31-(0)521-595255 fax:+31-(0)521-597332
> **************************************************************************
>
>
>

**************************************************************************
Ramesh K
(Mo/Tu): Radiosterrenwacht Westerbork, Schattenburg 1, 9433 TA Zwiggelte.
ph: +31-(0)593-598755 fax:+31-(0)593-592486
(We/Th/Fr): Stitching Astron, Oude Hoogeveensedijk 4, 7991 PD Dwingeloo.
ph: +31-(0)521-595255 fax:+31-(0)521-597332
**************************************************************************