On 08/03/2014 07:40 PM, Turbo Fredriksson wrote: > 1. Because there's no fsck before mounting the filesystem. > > That's really ugly, and potentially dangerous. Yes, it will simply fail if > the fs > is dirty, but I'm sure there's occasions where mount doesn't discover > everything > and a fsck should have been done first. Also, it's just common practice... >
And that's because there is no need to do so every time. Imagine a box with 512 LUNs. What you end up asking to do is to explicitly run fsck every time the OS boots. Besides, if you really want your devices/mount points to be checked at every boot, you should specify so in /etc/fstab mount options. > 2. There's no '-O _netdev' to swapon. There is for mount, so the > Well. In theory, yes, you can have swap on a network block device. But that would be of not much use. And even if you have such a case, you can always cook something local to handle that. For Debian's generic init scripts, I don't see much room for it. > 3. A 'mount -a' (or even 'mount -a -O _netdev') will/might include unwanted > filesystem. Currently the code will skip filesystems of the type > > nfs|nfs4|smbfs|cifs|coda|ncp|ncpfs|ocfs2|gfs|ceph > I don't understand this one. Anything that we mark as "_netdev" will be accounted for. No ? > 4. No other init script (that I've seen) simply "mounts everything". All > scripts > have some control over what's mounted and what's not (including swapon). > We are not mounting everything. Only the ones the user explicitly asked for, in /etc/fstab > > > Doing some tests, this seems to be ok: > > fsck -a -M -T -s > mount -a -O _netdev > swapon -a > > I'm not sure I like the 'do on everything' part though.. I don't like the fsck included in it. I'll do some investigation and accordingly, include or drop it. -- Ritesh Raj Sarraf | http://people.debian.org/~rrs Debian - The Universal Operating System
signature.asc
Description: OpenPGP digital signature
