Backup to a qnap nas

I don't have a QNap NAS, so I can't tell exactly. But from other responses I got about QNap it seems like you need to replace the original SSHd with OpenSSH. Take a look on http://forum.qnap.com/viewtopic.php?f=11&t=70425

The original 'find' command is also not working with BIT. You need to install 'findutils' through Optware. Next BIT version 1.1.8 will provide an expert-option to add commands before every ssh command. This will make it easier to link 'find' to findutils instead of busybox. You can already try this from our testing PPA (sudo add-apt-repository ppa:bit-team/testing && sudo apt-get update && sudo apt-get upgrade).

The alternative in current version is to activate "PermitUserEnvironment yes" in your /etc/ssh/sshd_config on remote host and create ~/.ssh/environment. Run this command:
ssh user@host echo "PATH=/opt/bin:/opt/sbin:\$PATH" > ~/.ssh/environment

You need to restart remoteOpenSSH after this.

Regarding your other questions please read the man-page 'man backintime'.

-----------------------------------------------------------

I'm always happy to help. But with WTF's inside questions I'm loosing interests in doing so...

Thanks for your detailed describtion. Maybe this will help other QNap users.
But I will skip this. I hate my QNap ;) and will migrate someday to a own build NAS.
I won't invest to much time in such a bad technic. Shame on me because I payed for that. :D

What do you think about...
I could create the snapshots on the local drive.
And after that I can use my own rsync-script with the option "--hard-links" to synchronize the complete snapshot directory to my nas.

I am still testing this and see no problems. But the process is to complex to be sure about that. So it would be nice to have your second opinon as a backup-professional.

Thanks for the credits, but I'm by far no backup-professional. I'm just an ordinary BIT-user who missed some function (the SSH mode).

It should work. But it will take ages when you have lots of snapshots and remote rsync could run out of RAM, too. Rsync would need to build an index for every snapshot. On my NAS it takes ~6-8min for just one snapshot. I have 74 snapshots so this would take ~7.5h without even a single byte of syncing.

If you don't want to fiddle around with you NAS's SSH I think it would be better to mount the NAS with NFS and sync to that local mountpoint. You can create a user-callback script which calls the mount command on $3=='7' and unmount on $3=='8'. Or you can use autofs to mount the share automatically.

This is nearly the same like my own-build/ugly backup-script work. It mounts a Samba-share into my local file-system and than use on my local machine rsync to backup to there.

Your solution sounds great.
But what is if the NAS is not on power (it doesn't run 24/7)? What choices do I have then?
Can tell inside the callback-script to give a message/sound and then wait (e.g. 10min) and try again?

Or will BIT just try to backup to a unmounted directory, give an error message about it and try it again?
How does BIT react if a backup doesn't succeed? I see no setup options for that.

How BIT react if an error occured was a question of mine in the past. But I can not find it again here on Launchpad.
LP is so bad I can not find my own question. I am not able to edit my own postings, ...

In the man page I can not find something relevant about "retry" or "error". Maybe there is no documentation about that very important behaviour or I can not find it. You see again the docu-problem? :)

I think contributing to the documentation would be nice for me. But before that we have be clear about the concept and how is really in charge for BIT.

BIT will check for the backintime/HOST/USER/PROFILE_ID folders. If they are not present it will show a message and wait 30sec for them to come up (I know, it's not yet configurable). You could use user-callback $3=='1' to do any other action you would like to. E.g. give a sound and custom message, wait for 10min and exit 1 if those folders are still not available. BIT will stop taking a new snapshot if the user-callback $3=='1' returned none zero.

There is no real error handling yet. But if you use the 'Repeatedly (anacron)' schedule, BIT will not write a new timestamp if there was an error and just try again after 15min. With the other schedules it will just start next scheduled time again (I know, that's no good).

-----

I don't like Launchpad either. I'm already thinking about moving BIT to Github. But I'm a bit afraid of all the consequences.

I'd highly appreciate contributions for documentation. And of course I would help on that by adding my knowledge and text blocks. But at the moment there are so many other things that need my attention in BIT code that I can't lift a complete new documentation by my own.

I think, it would be best to ship offline docs with BIT instead of a new website which will be down in couple of years again. Dan started this once in http://bazaar.launchpad.net/~bit-team/backintime/trunk/view/head:/qt4/docbook/en/index.docbook

What do you mean with 'in charge'? In my mind there is no 'in charge' concept in FOSS. There are just contributors. Dan invented BIT but left the project. Now I have control over the Launchpad project page and I'm at the moment the only one who is continuously working on it.

This question was expired because it remained in the 'Open' state without activity for the last 15 days.