Date: Sat, 26 Nov 2011 16:19:23 +0100 From: Ivan Voras <ivoras@freebsd.org> To: Kirk McKusick <mckusick@mckusick.com> Cc: freebsd-fs@freebsd.org Subject: Re: Does UFS2 send BIO_FLUSH to GEOM when update metadata (with softupdates)? Message-ID: <CAF-QHFVw9hXdN=X%2BM4kHXVRp=22Dc4zf6OJdJni6yh6vp_WGrQ@mail.gmail.com> In-Reply-To: <201111260727.pAQ7Rlh6056867@chez.mckusick.com> References: <jals5e$96f$1@dough.gmane.org> <201111260727.pAQ7Rlh6056867@chez.mckusick.com>
next in thread | previous in thread | raw e-mail | index | archive | help
On 26 November 2011 08:27, Kirk McKusick <mckusick@mckusick.com> wrote: > Your proposed change is definitely useful, though really just applies > to filesystems running without SU and SU+J. The cases where bwrite > are used are when we are using synchronous write to maintain > filesystem consistency (e.g., the filesystem before SU). Ok but I think the question here is: are synchronous writes enough, or a BIO_FLUSH needs to be sent down to the drives at critical moments? You are the best to say what is good for UFS but from what I've learned from other filesystems (including Linux), it seems that all of those use some kind of write barrier / BIO_FLUSH to be safe. Some docs: http://docs.redhat.com/docs/en-US/Red_Hat_Enterprise_Linux/6/html/Storage_Administration_Guide/writebarr.html
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?CAF-QHFVw9hXdN=X%2BM4kHXVRp=22Dc4zf6OJdJni6yh6vp_WGrQ>