Comment # 13 on bug 1222429 from Wenruo Qu

(In reply to Jiri Slaby from comment #10)
> (In reply to Wenruo Qu from comment #9)
> > As a quick fix, just moving btrfs out of "NORECOVERY_FS" would be sufficient
> > to me.
> 
> But that would replay the log on a mounted filesystem which would brick it
> for the mounter, right? Or what am I missing? (I assume/understand that ro
> replays the log.)
> 
> AIUI, there is a cluster, someone has mounted a FS on it. There there is
> this installer (yast) which wants to probe all filesystems on the cluster,
> but not actually touch/write them, because they are mounted elsewhere.

Oh, I got the point, it's SAN, so there is nothing preventing double mounting
from different clients.

In that case it's true that we can not do any write, including log replay.

Then we need to maintain a per-fs mount options to prevent log replaying.
We choose not to adapt "norecovery" name, because it takes me quite several
minutes to
remember it's no log replay... Other than my initial idea to disable rescue
mount options.