Date: Wed, 1 Jul 2009 13:29:44 -0400 (EDT) From: Rick Macklem <rmacklem@uoguelph.ca> To: "Julian H. Stacey" <jhs@berklix.com> Cc: Kirk McKusick <mckusick@mckusick.com>, Attilio Rao <attilio@freebsd.org>, freebsd-current@freebsd.org, freebsd-fs@freebsd.org Subject: Re: umount -f implementation Message-ID: <Pine.GSO.4.63.0907011319320.16855@muncher.cs.uoguelph.ca> In-Reply-To: <200907010048.n610mPem058027@fire.js.berklix.net> References: <200907010048.n610mPem058027@fire.js.berklix.net>
next in thread | previous in thread | raw e-mail | index | archive | help
On Wed, 1 Jul 2009, Julian H. Stacey wrote:
> Kirk McKusick wrote:
>> forced unmounts. The gentle force (-f) and the brute force (-F)
>> unmount.
>
> -F Would also be nice for devd.conf detach, for when people
> forget & pull a USB stick without unmounting first.
> Better a corrupt stick than a crashed OS.
>
All I'll add is, for the experimental nfs client, if both semantics are
desired (and, imho they are), there will need to be separate flags to
indicate whether or not to terminate RPCs in progress.
So, it seems that there is interest in a separate "umount -F" to handle
the case of failed storage (disk subsystem, NAS server down,...).
Is there anyone who is opposed to my pursuing this after FreeBSD-CURRENT
branches from 8? (I can do the experimental nfs client + some testing.
Hopefully others can help with the generic VFS issues and other file
systems.)
rick, who obviously doesn't have as good a memory as Kirk's:-)
ps: Unfortunately Solaris uses "-F" for something entirely different, so
feel free to suggest other flag values if you think that is a concern.
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?Pine.GSO.4.63.0907011319320.16855>