How would you best recommend backing up our seafile data? We have many large files, total seafile-data folder is 3.3TB. They are not encrypted, so I was thinking mount with FUSE and copy to a backup drive once a week. But I need to find a way to automate this as well as keep a few revisions of the backup.
Dump your database, make a snapshot and copy data + database backup + configuration files to the external drive. Once done do a snapshot there. Use rsync with delete option (only new and deleted files need to be transferred to the external hdd) and delete the snapshots if you have enough versions available on the external disk.
There is already a topic with some good ideas: Backup strategy for Seafile CE
Personally I’m quite satisfied with rsnapshot (
sudo apt-get install rsnapshot). It can also keep revisions.
That sounds complicated…
How does rsnapshot work? I’m not familiar but I looked at their page. But can you tell me basically how it works?
It mentioned to use the How-To, but it dont exist. So I am reading through the FAQ right now.
We run seafile on a debian VM, could something like this do the trick for us?
It is actually very easy.
On Ubuntu 16.04, I am using Btrfs snapshots and btrbk for backup to a different disk or remote location. https://github.com/digint/btrbk
sorry for the late reply. rsnapshot actually uses the rock solid rsync-tool. Once installed you configure it in this file:
sudo nano /etc/rsnapshot.conf. Here you mainly add a backup root, for example
snapshot_root /media/backup_usb/rsnapshot/and backup points for example your seafile folder
backup /home/seafile/ seafile-data_backup/.
- If you use a external drive for your backups make sure to set
no_create_root 1, so backups are only done with the drive connected.
- After editing check that the syntax of your configuration is correct with:
sudo rsnapshot configtest
- Start your first backup with:
sudo snapshot daily
- For automated backups configure cron. For example add these lines to your cron file:
sudo crontab -e
30 3 * * * /usr/bin/rsnapshot daily >> /tmp/rsnapshot.log 2>&1
30 6 * * 0 /usr/bin/rsnapshot weekly >> /tmp/rsnapshot.log 2>&1
30 8 1 * * /usr/bin/rsnapshot monthly >> /tmp/rsnapshot.log 2>&1
yes it could. Veeam is a good choice generally. You want to backup from hypervisor level or machine level?
Veeam linux backup agent is pretty new and might still be buggy. You might be better off with the Acronis Linux backup agent.
Very interesting, I use rsync so far. Very simple backup script which does a DB dumb and then triggers an rsync to a CIFS share. No snapshots yet, which makes your solution quiet interessting.
Maybe we can work on a working script together that servs all in the community?
That would be great. I wouldn’t be much help myself though.
Have you used Veeam to backup VM’s? I would need to back up whatever way is most efficient. I have 4 VM’s a windows server and 3 linux. I need a way to automate backing it all up.
We use Veeam to Backup our Hyper-V VMs. It works really nice!
What Hypervisor do you use?
We use vmsphere client with esxi server
Veeam Backup and Replication is a very good tool, but also expensive. Call them and ask for an offer.
That one comes close to VMware prices and is usually licensed by sockets.
I’m going to try the free version. Not sure if a limitation is maybe I’ll have to manually run it or what though. If thats the case its not a huge deal, I’ll just run it once a week.
For ESXi backups you can use XSIBACKUP PRO https://33hops.com/xsibackup-pro-vmware-esxi-backup.html and this integrates with Borg Backups https://borgbackup.readthedocs.io/en/stable/ for VM or data snapshots.
i can recommend syncthing if you want to backup your data off site instead of rsync(-alike tools). As filesystem for the seafile node i can recommend zfs. Works perfectly fine and gives snapshot capabilities.
I tried a few weeks ago veeam backup tool for home use. Unfortunately it lacked on zfs support.