incremental backups
in [UK Linux]
|
Prev: BIOS 'Hanging'
Next: Text encryption
From: Andy Botterill on 5 Jun 2010 16:26 On 06/05/2010 08:54 PM, Dave Liquorice wrote: > On Sat, 05 Jun 2010 14:57:35 GMT, unruh wrote: > >>> Does unmounting effectively disable the usb disk (err powering it >>> down)? >> >> No, almost certainly not. USB power is continuous ( you can plug in a >> flashlight). > > True enough, the OP needs to define what he means by "powering it > down". Does he mean just the disc spun down or all the electronics > off as well. I mean the disc spun down. There appears to be a windows utility to do that. I do see a few websites which describe regular unix ways to do this. > The former might happen automagically or might be a config option in > the particular USB drive or not at all. It is also said that if no accesses occur within 10 minutes it automatically shuts down. At the moment my script is saying most of my personal directory has changed. I've done something wrong here. I need to fix that before trying spinning it down. Thanks for the hints. Andy >
From: Martin Gregorie on 5 Jun 2010 16:44 On Sat, 05 Jun 2010 14:57:35 +0000, unruh wrote: > On 2010-06-05, Andy Botterill <andy(a)plymouth2.demon.co.uk> wrote: >> I'm writing some scripts to do full and incremental backups for my >> computer. My incremental (daily) backup works. However yesterday the >> daily backup was 44GB ( I would say 1.4GB is normal but high ). > > Use rsync or rsnapshot to do your backing up. > Agreed. rsync is a very fast way of making a backup once the initial copy has been made. If you want to retain several backup generations (say a set of three daily backups). This is easy with rsync because it can make a whole filesystem backup into a specified target directory - simply rotate your daily backup round a set of root directories. > No, almost certainly not. USB power is continuous ( you can plug in a > flashlight). > I think this depends on the USB drive's controller have an old, WD 3.5" USB drive that powers off on a time-out after its unmounted. I also have a Formac 2.5" drive that doesn't. -- martin@ | Martin Gregorie gregorie. | Essex, UK org |
From: Gordon Henderson on 5 Jun 2010 17:26 In article <87631xkevu.fsf(a)araminta.anjou.terraraq.org.uk>, Richard Kettlewell <rjk(a)greenend.org.uk> wrote: >Andy Botterill <andy(a)plymouth2.demon.co.uk> writes: > >> I'm writing some scripts to do full and incremental backups for my >> computer. My incremental (daily) backup works. However yesterday the >> daily backup was 44GB ( I would say 1.4GB is normal but high ). >> >> Looking at the find instruction I have used -mtime -ctime and -atime. >> http://www.brandonhutchinson.com/ctime_atime_mtime.html >> The above article implies that I don't really need to monitor -atime. >> Is this a safe/sensible thing to do? > >atime is the time of the last read; you don't need to make a new backup >just because atime has changed. > >Personally I often disable atime completely. There are still one or 2 things that need it - at least I think there is (other than backup programs that is - and I think you need to use something like 'dump' so as not to disturb the atime when you read the file to archive it). Thinks like email programs I think.. >> The backup medium is a usb external drive. Eventually I would like to >> only mount the drive for the backup process. Does unmounting >> effectively disable the usb disk (err powering it down)? > >It won't power it down. > >I do backups to external drives too; however rather than monitoring >timestamps to construct incrementals, I used rsync with the --link-dest >option, so my backups are complete filesystem snapshots, with unchanged >files connected as hard links. Ah, that must be new (relatively speaking). I've been using cp -al .... followed by rsync for several years now... maybe that's faster. Must check! Gordon
From: unruh on 5 Jun 2010 20:37 On 2010-06-05, Bernard Peek <bap(a)shrdlu.com> wrote: > On 05/06/10 20:54, Dave Liquorice wrote: >> On Sat, 05 Jun 2010 14:57:35 GMT, unruh wrote: >> >>>> Does unmounting effectively disable the usb disk (err powering it >>>> down)? >>> >>> No, almost certainly not. USB power is continuous ( you can plug in a >>> flashlight). >> >> True enough, the OP needs to define what he means by "powering it >> down". Does he mean just the disc spun down or all the electronics >> off as well. >> The former might happen automagically or might be a config option in >> the particular USB drive or not at all. > > It would be really useful to be able to switch a USB port's power off > under software control. It might be. It isn't. Unplug the apparatus. > > >
From: Tom Anderson on 6 Jun 2010 08:31
On Sat, 5 Jun 2010, Gordon Henderson wrote: > In article <87631xkevu.fsf(a)araminta.anjou.terraraq.org.uk>, > Richard Kettlewell <rjk(a)greenend.org.uk> wrote: >> Andy Botterill <andy(a)plymouth2.demon.co.uk> writes: >> >>> I'm writing some scripts to do full and incremental backups for my >>> computer. My incremental (daily) backup works. However yesterday the >>> daily backup was 44GB ( I would say 1.4GB is normal but high ). >>> >>> Looking at the find instruction I have used -mtime -ctime and -atime. >>> http://www.brandonhutchinson.com/ctime_atime_mtime.html >>> The above article implies that I don't really need to monitor -atime. >>> Is this a safe/sensible thing to do? >> >> atime is the time of the last read; you don't need to make a new backup >> just because atime has changed. >> >> Personally I often disable atime completely. > > There are still one or 2 things that need it - at least I think there > is (other than backup programs that is - and I think you need to use > something like 'dump' so as not to disturb the atime when you read the > file to archive it). Thinks like email programs I think.. There was a big discussion about this on the LKML a while ago, when it was suggested that atime should be off be default, and there was this relatime replacement for it. IIRC, it was observed that most systems run perfectly happily without atime. It's all buried in here: http://kerneltrap.org/node/14148 tom -- Everything that has transpired has done so according to my design. |