Daniel Persson <[EMAIL PROTECTED]> wrote:
>Sorry for posting this here and not on the rsync mailing list(im on enuff
>mailinglists as it is :)
I know the feeling. ;-)
>I am trying to set up a working rsync enviroment where i will make one
>machine a complete mirror of on other (so that if the "master" goes down
>all i need to do is to change ip adress and keep going (in theory).
>
>The problem is, i have built a simple script with a single command line
>that (for now) syncs the /tmp on the master to the slave.
>
>Im using SSH and im logging in as root - even tough this i get permission
>denied on some of the diretories (for example the one that Helix Gnome's
>install program created when i ran it as root).
>
>I know that the user specified in /etc/rsync.secrets dosent have anything
>to do with the real users, but since im logging in as root trough SSH i
>felt like i would have all the rights to sync whatever files i want ?
Sounds to me like you should have rights to whatever as well...but I
haven't used rsync in this manner. If data encryption isn't horribly
important you may want to consider running rsync as a server/daemon process
and updating the backup server that way instead of over ssh. I do just
that to keep some data in sync between a couple of servers. Instead of
/etc/rsync.secrets you use a config file - /etc/rsyncd.conf whose syntax is
very similar to samba - which should come as no surprise as it was written
by some of the same people. At any rate, you can setup "share" points with
read-only permissions and map the uid/gid as well, something like so:
[sharename]
path = /
uid = 0
gid = 0
read only = yes
hosts allow = myhost.mydomain
I then use a simple shell script run by cron to keep things in sync.
>Is Rsync the best suited for this ? (i can live with a couple of minutes
>of downtime - not more)
Depending on the hardware, you may have other options. I have a couple of
HP NetServer's with front-mounted hot-swap drives. One server has a
"spare" boot disk kept in sync with rsync and if the first disk fails, I
down the server, remove the bad disk, but the good disk in its place and
reboot. (There are a couple more steps, but that's essentially it.)
The second server also uses the same drives so I could move the boot/data
disks to it if say the CPU or motherboard or whatever went 20 toes up on
the first.
Either of these may or may not be possible in you situation, or you may
feel this option requires too much work or too much downtime...but they are
other options you can consider.
>I looked at The Linux HA project - but it dosent seem finished ?
It isn't. I know the guy who's the lead engineer for the project and it's
coming along, but I don't think it's quite ready for public consumption yet.
>I also looked at Pirhana - and it seems to be primarly a load balance
>cluster - i need a failover cluster.
>
>I read the linuxworld article about the subject - and felt like there
>wasnt really any finished GPLed products out there.(at least not to run at
>customers that want HA.)
I don't think there are any that are quite serviceable as of
yet. TurboLinux supposedly has some type of failover-clustering available
with their distro, but I haven't seen it so I can't comment.
>Should i look into GFS or anything like that ?
Don't even know what it is...
Hope this gives you a few ideas, let me know if you need help with running
rsync as a daemon - there really isn't much to it.
-Eric
Eric Sisler
Library Computer Technician
Westminster Public Library
Westminster, CO, USA
[EMAIL PROTECTED]
Linux - don't fear the Penguin.
Want to know what we use Linux for?
Visit http://gromit.westminster.lib.co.us/linux
--
To unsubscribe: mail [EMAIL PROTECTED] with "unsubscribe"
as the Subject.
