Migrating from SME 9.2 - nethserver-hotsync

Try it first without customization and work from there ?

I’ll check and report back.

@mark_nl

Unfortunately no difference. Even with a systemctl restart rsyncd.

Yes the problem is the /usr sync I think. I suggest you to reinstall the slave and not backup these folders:

/etc, /lib, /usr

If you don’t have any customizations.
/usr folder contains hotsync files.

Does it make sense to exclude in /usr the hotsync folder rather then not backing up the whole folder?

Do you mean the whole server?

I think is better

You should backup default without adding folders. Add only custom folders (like in opt if you have any custom software)

O.k. I see.

For a quick test I excluded in /etc and /usr/sbin anything related to hotsync. Re-installed and re-configured the slave. It this doesn’t help I’ll re-install the server.

But why you add these folders to backup?

I installed a couple of non nethserver contribs. I.e. mailman and firebird-superserver. I wanted to make sure that they are also in the backup.

You should backup their specific folders, like /etc/example/firebird and not all /etc. Idem for other folders

1 Like

@federico.ballarini

I installed the backup-server in proxmox as a guest. Therefore I started from scratch. I followed your advise and included the exact location of the programs I installed on top of the NS. I issued hotsync on the master and hotsync-slave on the slave. No errors occurred. No emails from cron so far. After a reboot of the slave there were a few errors while booting up. I assume this is not relevant as log as not promoting the slave. I’ll keep an eye on it the next days. Finally thank’s again for your help on solving the issues.

regards,
stefan

2 Likes

If you want post here logs error and we can check if it’s normal

1 Like

I will do.

Question: the master is an PD AD, the slave isn’t configured yet. Just hotsync and hotsync-slave have been issued. Looking in cockpit there is i.e. now as Hostname/alias instead of before running the command hotsync-slave now just “1”. Before there was “backup.xxx.de”. BTW on the landing-page https://192.XXX.XXX:9090 there’s still the FQHN. Should I ignore this?

And - does it make sense before running hotsync-slave to configure the PD AD? In case of a promote wil it be restored on the slave? It’s not quite clear to me, because it has to be a free IP for the not yet promoted AD server? Could this be changed after a promote without loosing data? Maybe I’ll try in a VE…

No. On the hotsync slave you need only to configure hotsync parameters and then all your server will be restored with hotsync-promote action. Do not make configurations on slave machine.

Yes, I suggest you to try in VE

Sincerly I don’t now. If you logout login in Cockpit still remains 1? This is not normal.

As I configured a NS server in proxmox VE I rolled back to a vanilla point. Means before the command hotsync-slave. I issued the command on the slave. No errors occured. While checking the backup I’m missing directories I included in the system-backup. The reason may be the rollback of the slave. If the master “thinks” that the directories have already been synced to the slave, then this would be the reason.

I’ll rename a directory on the master and we’ll see if this is going to be synced to the slave. But if so - if I wipe hotsync from the master and the slave and re-install it, would that help for the whole process?

The engine is rsync, so if the directory does not exists or some files are missing, they should be copied.

I ended up in installing the slave new. I’ll let you know.

1 Like