xfs
[Top] [All Lists]

streaming media and realtime subvolume

To: linux-xfs@xxxxxxxxxxx
Subject: streaming media and realtime subvolume
From: George Georgalis <georgw@xxxxxxxxx>
Date: Thu, 7 Nov 2002 10:17:29 -0500
Sender: linux-xfs-bounce@xxxxxxxxxxx
User-agent: Mutt/1.3.28i
Hi -

I'm new to xfs and this list. Thanks to the kick that came from a HD
failure, I'm now running Debian, with my root, var and usr partitions 
on xfs. I used "blade's" debian install floppies.
http://people.debian.org/~blade/XFS-Install/

I was not happy with the default kernel. The original xfs kernel seemed
really fast, extracting tar files etc, but would hang altogether for
3 to 10 seconds at times (presumably in journal io). I made a new
kernel with xfs-1.1-2.4.18-all.patch.bz2 and linux-2.4.18.tar.bz2 and
everything seems to be running great.


I do have a few questions though. I'm not sure if blade's disks used
xfs-1.0 or xfs-1.1; and xfs-1.2 seems eminent (I saw some xfs-1.2preX
patches). Does the filesystem remain the same through these version 
changes or should I think about doing a mkfs.xfs each time I upgrade
the the xfs version?


One main reason I'm experimenting with xfs is its performance rating.
I'm designing a dedicated i386, IDE software RAID 1, media streaming
machine (whew). The box's only function will be to stream media from
its disk. This is actually a node of which there will be many, with a
central server.

I expect a single 7200rpm ide drive to be able to handle the stream.
I'm using raid for high availability; but uninterrupted streams is
a must. What mount parameters can I use to assure uninterrupted io?
Are there other adjustments I should be thinking of? I've seen a few
mentions of the Linux xfs realtime subvolume but no doc, is it ready for
production? From what I can tell, it's a non journeled contiguous data
block. Maybe I should just use ext2 for the media files? Would that be
higher performance? I'm not too worried about fsck, because in the case
of corruption I can remake the filesystem (data partition) and renew the
data from the node server.

Thanks for your input.

// George


-- 
GEORGE GEORGALIS, System Admin/Architect    cell: 347-451-8229 
Security Services, Web, Mail,            mailto:george@xxxxxxxxx 
File, Print, DB and DNS Servers.       http://www.galis.org/george 


<Prev in Thread] Current Thread [Next in Thread>