xfs
[Top] [All Lists]

Re: [PATCH 58/71] xfs: garbage collect old cowextsz reservations

To: Christoph Hellwig <hch@xxxxxxxxxxxxx>
Subject: Re: [PATCH 58/71] xfs: garbage collect old cowextsz reservations
From: "Darrick J. Wong" <darrick.wong@xxxxxxxxxx>
Date: Tue, 27 Sep 2016 12:29:19 -0700
Cc: david@xxxxxxxxxxxxx, linux-xfs@xxxxxxxxxxxxxxx, xfs@xxxxxxxxxxx
Delivered-to: xfs@xxxxxxxxxxx
In-reply-to: <20160927185021.GA8623@xxxxxxxxxxxxx>
References: <147216791538.867.12413509832420924168.stgit@xxxxxxxxxxxxxxxx> <147216829394.867.16281333542262043955.stgit@xxxxxxxxxxxxxxxx> <20160924194234.GA1878@xxxxxxxxxxxxx> <20160926215209.GE14092@xxxxxxxxxxxxxxxx> <20160927185021.GA8623@xxxxxxxxxxxxx>
User-agent: Mutt/1.5.24 (2015-08-30)
On Tue, Sep 27, 2016 at 11:50:21AM -0700, Christoph Hellwig wrote:
> On Mon, Sep 26, 2016 at 02:52:09PM -0700, Darrick J. Wong wrote:
> > The two functions /could/ be merged but I'm hesitant to combine them
> > because they run as different workqueue items.
> > 
> > Setting speculative_cow_prealloc_lifetime to a (much) higher value than
> > speculative_prealloc_lifetime has been useful for combatting CoW
> > fragmentation on VM hosts where the VMs experience bursty write
> > behaviors and we can keep the utilization ratios low enough that we
> > don't start to run out of space.  IOWs, it benefits us to keep the CoW
> > fork reservations around for as long as we can unless we run out of
> > blocks or hit inode reclaim.
> 
> Ok, so there is a good use case for it.  It just felt to me like
> there was a little bit too much duplication, that's why I asked.

I'll add a brief explanation of this to the patch changelog to
record the justification.

--D

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