[Top] [All Lists]

Re: hole punching performance

To: Dave Chinner <david@xxxxxxxxxxxxx>
Subject: Re: hole punching performance
From: "Bradley C. Kuszmaul" <kuszmaul@xxxxxxxxx>
Date: Tue, 15 Jan 2013 21:46:36 -0500
Cc: Florian Weimer <fw@xxxxxxxxxxxxx>, Ben Myers <bpm@xxxxxxx>, xfs@xxxxxxxxxxx
Delivered-to: xfs@xxxxxxxxxxx
Dkim-signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=20120113; h=x-received:mime-version:in-reply-to:references:from:date:message-id :subject:to:cc:content-type; bh=dDP3sEiuGbK1jBApmppGzu8JrpdpwwL7zDZRvYSPCqE=; b=zBM2vt4VATMOMrIFMZUGQfirhxXWEyJL0y3r5QuD/N3ahl40h5JAzD3AlWh0bvaMG7 mmgBLcsfpdTedi6NXyd5ETV11xclH+PuYMvtUWFRy0np3GSJEAsAgDLtQ6F7cnbTqM5B dRW0pefCeqCaIf13UAi7acC4NNwQAPyS6y28swJJDX7n75sEeyDKNQUqUmxutM9KT/YV JkvlBowEs4CeQK6xEuLHY/kSn+plU5VF8c0Hh3zSHsDy2+365/acTQK+EiAKOhevD83+ 0s6C07KbH2FG4+pTTuCSCI6Yv0Akbd2JkdexbLUpjPbGEoc4qdaDogY9LuBrKAemLl2o L/nw==
In-reply-to: <20130116000051.GC2498@dastard>
References: <CAKSyJXf66H2U-BF-aYnSr2fF24_6LJw6swOx1RhUc_3Eqayaiw@xxxxxxxxxxxxxx> <87pq195t8q.fsf@xxxxxxxxxxxxxxxxx> <CAKSyJXfQsmemZ3H9A92UezsPCXfn37LXP3tpPMkiwtvR4=eY2g@xxxxxxxxxxxxxx> <20130113194941.GB27055@xxxxxxx> <20130113225545.GA19173@dastard> <87mwwaqgkj.fsf@xxxxxxxxxxxxxxxxx> <CAKSyJXdejNhNgqq6g8N4n6Cjz2s3+m=JUvanTRfDtF72d2RxFg@xxxxxxxxxxxxxx> <20130115225435.GB2498@dastard> <CAKSyJXfs-vfUKybDiQR2mSdP18am06CHsx0io6bF9HLWmvWoFA@xxxxxxxxxxxxxx> <20130116000051.GC2498@dastard>
But if I populate the file with a collection of near-1MB writes at random offsets, I'm likely to end up with one extent per write.  Each non-hole line in xfs_bmap is an extent, right?


On Tue, Jan 15, 2013 at 7:00 PM, Dave Chinner <david@xxxxxxxxxxxxx> wrote:
On Tue, Jan 15, 2013 at 06:33:54PM -0500, Bradley C. Kuszmaul wrote:
> So my understanding is that the first time I do a read() I'll have to read
> the extents.  My writes are on the order of a megabyte, a 250MB file will
> likely require an extra I/O to open.  That's fine.

write size != extent size.

Extents on XFS on a 4k block size filesystem can be up to 8GB in
size. It's entirely likely that your 250MB file is a single extent
an dhence has no extra overhead on read.  Use xfs_bmap to check.


Dave Chinner

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