At Sun, 21 Jun 2009 15:06:35 +0200 CentOS mailing list <centos@xxxxxxxxxx> wrote:
>
>
>
> I think about using dd instead of dump? Is this an acceptable idea?
*NO!*. dd is NOT a proper backup tool... I don't know where the idea
that it is comes from (probably some really old UNIX sys admin book or
something). dd has many uses, but it is not normally considered a
backup tool. Dump is a perfectly acceptable backup tool. Dump was
*designed* as the backup tool of choice. And it works just fine on a
read-write file system, with the *standard* disclaimer that if used on
a file system that is 'active', it (*like all backup tools*) will miss
files that are being created/writen to during the backup. Generally,
the missed files will be gotten on a later backup. On a normal, typical
system, one would run the full backup at a 'quiet' time (a time frame
of low activity, with the idea of avoiding having the backup interfere
with normal activities and of normal activities interfering with the
backup. On a typical normal system, the missed files will likely be
things like the tail end of log files or various trasienent files (like
mail queue files and the like).
If you want a non-live mirror on the second does do this:
Pick a 'quiet' time (say on a quiet Sunday morning) and do this:
(Assume that /dev/sdb1 is the sole partition on the second disk):
# mkfs.ext3 -L SecondDisk /dev/sdb1
# mkdir /seconddisk
# mount -v -t ext3 LABEL=SecondDisk /seconddisk
# dump 0f - / | (cd /seconddisk; restore -rf -)
At this point pick out a good book to read and get comfortable and read
a chapter or three or you can do whatever you would do to kill some time
-- ie go for a walk (or walk the dog), play a game of hoops or go for a
swim or whatever. I am assuming that the disk is probably good sized,
so full dump may take some time. You could stare at the screen for an
hour, if that floats your boat... Dump will display a progress report
every 5 minutes.
# umount -v /seconddisk
Now create a script (Let's call it '/usr/local/sbin/dailybackup'):
#!/bin/sh
/sbin/e2fsck -C -T -a LABEL=SecondDisk
mount -v -t ext3 LABEL=SecondDisk /seconddisk
rsync -v -a -x -H --delete --delete-after --exclude=lost+found/ / /seconddisk
umount -v /seconddisk
# chmod +x /usr/local/sbin/dailybackup
Now create a daily cron job:
# crontab -e
Add the line:
10 0 * * * /usr/local/sbin/dailybackup
And you are all set. Once a day an 10 past midnight, the backup disk
will be sync'ed to the live system disk. Every morning you will get a
message from cron with the output.
If you really want to, you can change the '0' above to '0,6,12,18' and
the sync'ing will happen every 6 hours. This is *probably* overkill.
>
> 2009/6/21 Robert Heller <heller@xxxxxxxxxxxx>
>
> > At Sun, 21 Jun 2009 11:49:09 +0200 CentOS mailing list <centos@xxxxxxxxxx>
> > wrote:
> >
> > >
> > >
> > >
> > > Hi all. I'm currently having a following problem: I have only ssh
> > connection
> > > to a CentOS 5.2 system, there are two harddiscs on it. One stores the
> > system
> > > (/ filesystem) and the other should be used to help restore the system in
> > > case of first disks' failure. I thought that maybe dump would be a good
> > > utility to make it. But in only works on read-only filesystems. In one
> >
> > Dump works just fine on a read-write file system. There is the pretty
> > much standard limitation (that applies to *all* backup methods) that
> > when backing up an 'active' file system: there will always be files that
> > will miss the backup because they were being written during the backup
> > process.
> >
> > > article I've read that making a snapshot of the / filesystem (then it
> > > wouldbe read-only) and backing it could help. But aren't snapshots
> > limited
> > > to logical volumes (LVM)? My friend told me to use rsync to back up the
> > > entire / filesystem to the second disk and then in case o failure the
> > system
> > > from the copy should boot ok.
> > >
> > > Could anyone provide any suggestions? I don't have physical contact with
> > the
> > > machine so for example RAID 1 isn't a possible option/
> > >
> > > Any help will be very kindly appreciated.
> >
> > Make an initial dump to get the base system copied, then set up a cron
> > job to sync the disks once a day (or more frequently) with rsync.
> >
> > >
> > > With regards,
> > > R.
> > >
> > > MIME-Version: 1.0
> > >
> > > _______________________________________________
> > > CentOS mailing list
> > > CentOS@xxxxxxxxxx
> > > http://lists.centos.org/mailman/listinfo/centos
> > >
> > >
> >
> > --
> > Robert Heller -- 978-544-6933
> > Deepwoods Software -- Download the Model Railroad System
> > http://www.deepsoft.com/ -- Binaries for Linux and MS-Windows
> > heller@xxxxxxxxxxxx -- http://www.deepsoft.com/ModelRailroadSystem/
> >
> > _______________________________________________
> > CentOS mailing list
> > CentOS@xxxxxxxxxx
> > http://lists.centos.org/mailman/listinfo/centos
> >
>
> MIME-Version: 1.0
>
> _______________________________________________
> CentOS mailing list
> CentOS@xxxxxxxxxx
> http://lists.centos.org/mailman/listinfo/centos
>
>
--
Robert Heller -- 978-544-6933
Deepwoods Software -- Download the Model Railroad System
http://www.deepsoft.com/ -- Binaries for Linux and MS-Windows
heller@xxxxxxxxxxxx -- http://www.deepsoft.com/ModelRailroadSystem/
_______________________________________________
CentOS mailing list
CentOS@xxxxxxxxxx
http://lists.centos.org/mailman/listinfo/centos
