[Gluster-devel] Segfault in read_ahead in 1.3.0_pre4

Harris Landgarten harrisl at lhjonline.com
Thu May 24 17:50:15 UTC 2007


For what its worth, stat-prefetch will likely be very important for the S3 translator as stat calls to S3 are very expensive. I have experimented with S3DFS to a 100,000 file bucket and neither find nor du ever return.


----- Original Message -----
From: "Anand Avati" <avati at zresearch.com>
To: "Brent A Nelson" <brent at phys.ufl.edu>
Cc: "Harris Landgarten" <harrisl at lhjonline.com>, gluster-devel at nongnu.org
Sent: Thursday, May 24, 2007 12:00:16 PM (GMT-0500) America/New_York
Subject: Re: [Gluster-devel] Segfault in read_ahead in 1.3.0_pre4

> I noticed that you are using stat-prefetch; have you encountered any
> issues? I was finding that du's on complex directories could return
> abnormal results and/or errors, so it seemed that heavy metadata queries
> were occassionally glitchy.  Without stat-prefetch, it's been fine.  If
> you've been having good luck with it, maybe I should try again.

Just FYI, stat-prefetch is likely to be discontinued in the
glusterfs--mainline--2.5 tree. the complexity and possible
inconsistancies is not worth the performance gain (helps only an ls -l
to be marginally faster). or we might just decide keep it for whatever
gains it gives (and to not break your config spec files) since you
always have the option of not loading it.

avati

-- 
Anand V. Avati






More information about the Gluster-devel mailing list