Date: Fri, 25 May 2012 08:24:11 -0500 From: Mark Felder <feld@feld.me> To: freebsd-fs@freebsd.org Subject: UFS SUJ and fsck questions Message-ID: <op.weu8uleu34t2sn@tech304>
next in thread | raw e-mail | index | archive | help
Hi guys,
I'm building out a highly available storage backend with HAST+NFS on UFS.
So far I've encountered a node going down and the filesystem being dirty,
so the other side won't mount it automatically. I've resolved this issue
with the following before the mount:
fsck -C -t ufs -y /dev/hast/${disk}
However, the problem is that on a disk nearly 1TB in size it will take a
long time to fsck and the failover won't be as smooth. SUJ would fit the
bill here pretty well. My main issue is not understanding the interaction
between SUJ and fsck.
If I simply try to mount a fs with SUJ, it will do the SUJ magic if
necessary and move on. But what if it's damaged beyond what SUJ can handle
and needs a real fsck? Can I use the same procedure? Will executing `fsck
-C` against an SUJ enabled filesystem that hasn't run the SUJ journal yet
do that first and exit if the journal replay was successful? If not, does
anyone have any ideas on how I can detect that automatically so I can get
the filesystem mounted cleanly without human intervention?
Thanks!
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?op.weu8uleu34t2sn>