On 02/24/2010 01:55 AM, [email protected] wrote: > hi! > > i've got two issues with autofs, first one: > > i want to mount an smb share. i have to pass username/password as > well as the server's ip as a mount option (plus uid, gid, and all > other smb options should be generally accessible using autofs, too, i > guess). the following doesn't work: > > /srv/smb > -fstype=smbfs,ip=10.1.2.3,user=administrator,password=secret,uid=smbuser,gid=smbgroup > //10.1.2.3/C$/
You can't start a mount location with "/" because it will be interpreted as a path instead of a mount location. The sun ":" escape has always been used to cover this case. eg. /srv/smb -fstype=smbfs,.... ://10.1.2.3/C$/ Not sure if you will have a problem with the trailing "/", you'll need to check. > > a mount command with these options (replacing the "-" by "-o ") > _does_ work. the funny thing is that such a manual mount won't work > using 10.1.2.3:/C$/ notation, obviously because that already makes > mount think the fstype must be nfs. > > what i already found out is that the //<server>/<share>/ notation > seems to cause problems to automount. using 10.1.2.3:/C$/ seems not > to cause the same problem as when using mount (automount seems to > still understand the fs type is smb) but it still fails, so it seems > to mess up the other options somehow. > > now that leads me to the second issue: > > the above fs is in the fstab and currently has to be mounted manually > when needed. now having that done by automount would be nice. that's > a simple direct mount, so the idea is to simply add to auto.master: > > /- /etc/auto.direct > > and to add to /etc/auto.direct: > > /srv/smb > > now the options and target are missing - on purpose. what i mean/find > most straightforward/find most intuitive, would be to have the > possibility to just add the mountpoint to the direct mounts, so that > automount will just mount the fs according to fstab (like when > calling mount manually with the mountpoint as the only argument). > that would btw. also fix/work around the issue #1 :) > > as i mentioned, to me this seems most intuitive. obviously - as > that's not (the way it was) implemented, others don't. but maybe it's > still regarded as a nice thing to have? That's quite an interesting idea and, to be honest, I hadn't thought of doing it that way. I'm not sure if that could lead to problems though so I'd need to think about it for a while. Ian _______________________________________________ autofs mailing list [email protected] http://linux.kernel.org/mailman/listinfo/autofs
