Re: [Server-devel] XS automount design
2009/8/3 Martin Langhoff : > The other thing that bothers me is that unmounting it automatically > makes a mess for anyone working interactively with USB sticks from the > commandline. > > Remounting it RO at least leaves the mountpoint in place, and it's > relatively easy and clear what to do. That would work. Daniel ___ Server-devel mailing list Server-devel@lists.laptop.org http://lists.laptop.org/listinfo/server-devel
Re: [Server-devel] XS automount design
On Sat, Aug 1, 2009 at 3:50 AM, Douglas Bagnall wrote: > Martin wrote: > >> - we don't have any scripts that write, but we will likely do, so >> mounting ro is not a good idea > > I'm pretty sure the xs-otp scripts write to the usb drive, if you have > that enabled. Not that anyone does that, or we'd have seen a stream of > queries about servers that are very hard to log into. True. I had forgotten - thanks! The other thing that bothers me is that unmounting it automatically makes a mess for anyone working interactively with USB sticks from the commandline. Remounting it RO at least leaves the mountpoint in place, and it's relatively easy and clear what to do. cheers, m -- martin.langh...@gmail.com mar...@laptop.org -- School Server Architect - ask interesting questions - don't get distracted with shiny stuff - working code first - http://wiki.laptop.org/go/User:Martinlanghoff ___ Server-devel mailing list Server-devel@lists.laptop.org http://lists.laptop.org/listinfo/server-devel
Re: [Server-devel] XS automount design
Martin wrote: > - we don't have any scripts that write, but we will likely do, so > mounting ro is not a good idea I'm pretty sure the xs-otp scripts write to the usb drive, if you have that enabled. Not that anyone does that, or we'd have seen a stream of queries about servers that are very hard to log into. Douglas ___ Server-devel mailing list Server-devel@lists.laptop.org http://lists.laptop.org/listinfo/server-devel
Re: [Server-devel] XS automount design
On Fri, Jul 31, 2009 at 2:03 AM, Daniel Drake wrote: > It works fine and the usbmount code is already crafted to cleanly > handle such a situation. how do you feel about this 98-umount mount.d > trigger? umount $UM_MOUNTPOINT && log notice "Unmounted" && exit log notice "Failed to unmount" exit 1 ...? m -- martin.langh...@gmail.com mar...@laptop.org -- School Server Architect - ask interesting questions - don't get distracted with shiny stuff - working code first - http://wiki.laptop.org/go/User:Martinlanghoff ___ Server-devel mailing list Server-devel@lists.laptop.org http://lists.laptop.org/listinfo/server-devel
Re: [Server-devel] XS automount design
2009/7/30 Martin Langhoff : > I don't know if usbmount is sane in the face of triggering the unmount > from the mount-post-processing scripts. It works fine and the usbmount code is already crafted to cleanly handle such a situation. how do you feel about this 98-umount mount.d trigger? Daniel 98-umount Description: Binary data ___ Server-devel mailing list Server-devel@lists.laptop.org http://lists.laptop.org/listinfo/server-devel
Re: [Server-devel] XS automount design
2009/7/30 Martin Langhoff : > No, we encourage them to wait for the "done" bells. Let me explain. > > - we don't have any scripts that write, but we will likely do, so > mounting ro is not a good idea > > - we're either mounting sync or with the "not-so-lazy async" that was > introduced a few kernels ago, so the sync is not delayed > > Are you getting the "I'm done" bells properly? Yes. However I don't think it's safe to unplug a rw-mounted filesystem at an arbitrary time, even if sync is enabled. We could either mount it ro, or we could mount rw but umount when done (just before the 2nd set of bells). > I don't know if usbmount is sane in the face of triggering the unmount > from the mount-post-processing scripts. I'll try :) Daniel ___ Server-devel mailing list Server-devel@lists.laptop.org http://lists.laptop.org/listinfo/server-devel
Re: [Server-devel] XS automount design
On Thu, Jul 30, 2009 at 8:31 AM, Daniel Drake wrote: > It seems a little strange to me that the XS scripts that process files > on USB sticks do not unmount the disk after use. In fact they don't > even mount the disk in read-only mode. > > Do we encourage users to login as root and unmount the disk before > unplug, or are we actually encouraging users that unmount is not > necessary? No, we encourage them to wait for the "done" bells. Let me explain. - we don't have any scripts that write, but we will likely do, so mounting ro is not a good idea - we're either mounting sync or with the "not-so-lazy async" that was introduced a few kernels ago, so the sync is not delayed Are you getting the "I'm done" bells properly? I don't know if usbmount is sane in the face of triggering the unmount from the mount-post-processing scripts. hth, m -- martin.langh...@gmail.com mar...@laptop.org -- School Server Architect - ask interesting questions - don't get distracted with shiny stuff - working code first - http://wiki.laptop.org/go/User:Martinlanghoff ___ Server-devel mailing list Server-devel@lists.laptop.org http://lists.laptop.org/listinfo/server-devel
[Server-devel] XS automount design
It seems a little strange to me that the XS scripts that process files on USB sticks do not unmount the disk after use. In fact they don't even mount the disk in read-only mode. Do we encourage users to login as root and unmount the disk before unplug, or are we actually encouraging users that unmount is not necessary? Daniel ___ Server-devel mailing list Server-devel@lists.laptop.org http://lists.laptop.org/listinfo/server-devel