[Top] [All Lists]

Re: mkfs.xfs with a 9TB realtime volume hangs

To: Eric Sandeen <sandeen@xxxxxxxxxxx>, Jan Wagner <jwagner@xxxxxxxxxxx>, xfs@xxxxxxxxxxx
Subject: Re: mkfs.xfs with a 9TB realtime volume hangs
From: Eric Sandeen <sandeen@xxxxxxxxxxx>
Date: Sun, 11 Jan 2009 07:46:24 -0600
In-reply-to: <20090111103528.GA8071@disturbed>
References: <alpine.DEB.2.00.0811141231390.23165@xxxxxxxxxxx> <49690F73.1070908@xxxxxxxxxxx> <20090111103528.GA8071@disturbed>
User-agent: Thunderbird (Macintosh/20081209)
Dave Chinner wrote:
> On Sat, Jan 10, 2009 at 03:13:23PM -0600, Eric Sandeen wrote:

>> With the default 4k extent size this takes forever (the man page claims
>> default is 64k, maybe this got broken at some point).
> It got changed a few years back by Nathan, IIRC. 

Yep, I found the commit yesterday, it was for buffered IO's benefit on
the rt subvol, to reduce unwritten extent conversion IIRC.

I was going to follow up w/ the commit etc yesterday, but the mailing
list was down (again, sigh - this is indicative of sgi's staunch
commitment to xfs, I'm sure) so had nothing to reply to.


> I bet the time
> being taken a result of the blow-out in bitmap size caused by reducing
> the extent size. Given it is non-linear, it may have something to do
> with cache sizes as well.  e.g buftarg hashes not large enough.
> Cheers,
> Dave.

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