If you have a tar or other backup you ought to be fine once they are restored. You just need to make sure the permissions on the files are set properly once you restore. For example, you do not want them owned by root or something. So just chown -R the whole thing to the user:group that you run your svnsync process as and you should be fine.
Mark
> On Apr 7, 2019, at 11:11 AM, Bo Berglund <bo.berglund_at_gmail.com> wrote:
>
> I have an Ubuntu Server 16.04.2 LTS that is running Apache Subversion
> 1.9.7 (r1800392). Total size of repositories is 5.5 GB. It was
> installed a year ago approximately.
> This server is acting as a backup to the main repository using
> svnsync.
>
> Today I was going to upgrade the Ubuntu server to version 18.04.2 as
> suggested when I logged in to the server (using SSH). So I was going
> to use the week-end to upgrade.
>
> But it went south on me, the do-release-upgrade process chugged along
> until it suddenly disappeared from my monitor. Turns out something
> went really bad with the upgrade, but all data files seem to be OK.
> Except when booting the server goes into PANIC mode.
>
> So I figured I would have to extract the disk data to an USB disk or
> similar, before installing Ubuntu Server 16.04.2LTS from scratch
> again.
> In the console which I can reach using an earlier kernel I can tar up
> the SVN repositories, so if I can just move them out of the PC as tgz
> files I would be OK.
>
> Once that has been done and Ubuntu is reinstalled I need to attach the
> files from the original disk to the svn server. How would that be best
> done?
>
> If I extract the archives into /var/lib/svn would svn be OK with the
> files then?
> Does it do some magic like marking the files in any way that will make
> the sync from the main server fail?
> Or is the ID of a repository stored in some of the files inside the
> repo database?
> The svnsync function relies in some way on the repository ID,
> right....
>
>
> --
> Bo Berglund
> Developer in Sweden
>
Received on 2019-04-07 17:27:30 CEST