Skip site navigation (1)Skip section navigation (2)
Date:      Sun, 20 Apr 2003 00:56:24 +0200
From:      Marko Zec <zec@tel.fer.hr>
To:        Chris Pressey <cpressey@catseye.mb.ca>, freebsd-stable@freebsd.org
Subject:   Re: PATCH: Forcible delaying of UFS (soft)updates
Message-ID:  <200304200056.24781.zec@tel.fer.hr>
In-Reply-To: <20030419172756.17aaf627.cpressey@catseye.mb.ca>
References:  <3E976EBD.C3E66EF8@tel.fer.hr> <3EA1B72D.B8B96268@mindspring.com> <20030419172756.17aaf627.cpressey@catseye.mb.ca>

next in thread | previous in thread | raw e-mail | index | archive | help
On Sunday 20 April 2003 00:27, Chris Pressey wrote:
> A simple comment to this effect (like 'all writes are delayed, so they
> go out to disk in a burst which can suspend the machine for a long
> time') next to the option in question would serve the same purpose.

I agree 100%. However, note that write bursts will typically occur combined 
with disk spinups. As it takes one to two seconds first to spin up the disk, 
the additional couple of _miliseconds_ will be completely unnoticable in 
almost all cases.

> Isn't the danger, basically, that a computer running FreeBSD with this
> patch could potentially:
>   - accumulate a LOT of future writes to be done all at once

Yes, it could accumulate a lot of writes to be done all at once - that is the 
whole purpose of the patch.
No, it will not accumulate more dirty buffers than the standard code, as the 
syncer will start writing the buffers to the disk once the resources become 
scarce.

>   - exhaust its resources when actually synchronizing those writes

No (I guess :)

>   - panic

Not a single panic until now.

> Terry: try to keep some perspective.
> Marko: clearly mark the patch 'experimental'.

Well, I'm affraid after this debate with Terry I can only mark the patch as 
"dead" :(

Marko



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?200304200056.24781.zec>