[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: dd or cp over network: should I use scp?





On Thu, Feb 24, 2011 at 4:33 PM, Celejar <celejar@gmail.com> wrote:
On Thu, 24 Feb 2011 15:48:00 -0500
shawn wilson <ag4ve.us@gmail.com> wrote:

> On Feb 24, 2011 3:40 PM, "Celejar" <celejar@gmail.com> wrote:
> >
> > On Thu, 24 Feb 2011 10:22:56 -0500
> > Curt Howland <Howland@priss.com> wrote:
> >
> > ...
> >
> > > But in general, it's necessary to back up only /etc and /home. These
> > > are where settings and user data are stored, and rebuilding the whole
> > > system it can be better to build the "system" anew, then just recover
> > > the user data and any needed custom settings from /etc.
> >
> > I'd strongly recommend backing up at least parts of /var (what if you
> > have mail in the mail spool when the system dies?).
> >
>
> I have a development vm that I backup with a simple svn commit from my home
> directory.
>
> In other words, your backup scheme should depend on what the system is used
> for.

Certainly - but my advice stands for typical, general purpose desktops,
laptops and servers.

disclaimer: be warned! in rereading this, i think i wrote more of a rant than anything informational. the 'backup everything' statements sorta got to me i suppose.

funny, i don't have a "general purpose" anything that needs all that backed up. the closest thing to 'general purpose' i have is graphic terminals - a macbook pro and a kubuntu box. however, for those machines, the only thing i really care about is my firefox profile. other than that, i have single purpose machines / vms - a mac mini media server (has one directory i care about), a dev box (backed up with svn), a mysql server (mysqldump does it there), drupal test (i'm not backing that up at all :) ), i backup the mini and sql server between each other and the sql server holds backups for everything else.

 so, what everyone is saying is that, on any and all linux computers, one should keep a backup of:
/etc - for configurations - why? this is pointless if i don't change anything / much. i generally end up shutting down unneeded services on new installs, but unless there's apache or slight changes to my.cnf (which i always have other machines with edited configs of) i don't really change much.
/home - my dev box has a /home/cpan which i rsync daily, do i really want to back that up? that's pointless imo. and, there's only one real user there - me!
/var - WTF? why? tons of useless old logs to delete? a mail spool full of cron telling me that things have been done? my databases that were backed up elsewhere? did i leave anything out or is this absolute stupidity?
/root - oh this is great, here is my crap. yes, full of wget tests or partial log files when i'm looking for something but i'm not exactly sure what it is. or, c programs that have already been installed (that i've probably documented and maybe even made an svn trunk for), i see a file called ./fail, ./grep, and ./generate.txt in there too. yeah, i'm sure i'll want to go through that pile of crap after rma'ing a hdd.
/usr/share/perl5 - now i have thought of backing this up (i don't). however, if shit hits the fan and i have to spend a day recompiling everything that i use (or other modules might use), i'm going to want to kill lots of things - probably every things. but, my other thought is that i don't do an update until there's a feature or something that i want, so it'd probably be good. i'd still want to kill though.

Reply to: