Whilst trying to do the upgrade to pro - I have managed to lose ALL of my files from Seafile.
There was around 2TB worth of data. and I have no backups of this (as I was going to start backing up once it was complete).
Please can someone help me get all of these back? (it looks like I can still see all of the blocks in the folder)
I recommend you do these these steps in this order. They won’t be enough for a full recovery, but will at least verify that your data isn’t completely gone.
Back up now. There’s a good chance your data still exists, but there is also a chance that if something goes wrong while trying to recover it could be lost. At least make a copy of the database and the seafile data. Backup and Recovery - Seafile Admin Manual
Try to use the seaf-fuse.sh script to mount your libraries and see the files they contain. If this works you can get your files. In the worst case you could just grab them from here and load them into new libraries after re-deploying your server, but it probably wont come to that. So for now unmount it and go on with troubleshooting. FUSE extension - Seafile Admin Manual
The experts here are probably going to need more details about what parts of the seafile system are working and what parts aren’t. I will help if I can, but hopefully someone more experienced will chime in when they have more to work with. Is there anywhere you can see your files (for example they show in the web interface but not in SeaDrive)? Are there any errors in the seafile logs (seafile, seahub, etc.)? Where there any errors while you were switching to the pro version?
Hi Tom,
Thank you so much for your response.
From what I can gather with it all, it’s just the database that’s been overwritten so it doesn’t contain the links - as the space used on my server is still around 2TB.
Which - from my understanding - means the Seafile-data folder still has the information in the blocks (a few new folders have been created which is why I presume this).
I have turned the server back on, stopped Seafile and seahub and I am now doing a back up of the Seafile data folder (found a guide that was reference v7.0…)
For reference, they aren’t showing in the web or the drive app, which is why I presume that it’s a database thing (as the ./setup-seafile-MySQL.sh file was run but was meant to connect to the existing DB)
No errors are appearing, just the files aren’t showing and I did have to re-create the admin account.
I will let you know an update of the seaf-fuse script mentioned after the backup is completed (doing that as a 100% security measure).
Thank you again for you help!
Kind regards,
Mitch
FYI to all, I think (hopefully) I may have got this right, If I have, I will post a full 0-100 guide on how to fix this in-case anyone has any issues with this in the future!
I believe that I have managed to fix the issue with a mixture of using SQL entries to force the old folders into the database, and running FSCK on the new instance.
I will post a full post regarding this as soon as possible, with the pre-requisites behind it as well.
Thank you for your response Tom and the Seafile Support team!
Trying to do the same crap after I selected do not sync and Seafile obliterated all of my files on all synchronized stations. When I try to download or view a single file from the snapshot in the web UI it doesn’t download and it doesn’t preview “failed to fetch”.
Really fun stuff. I feel like I trusted important stuff to some half-baked software made by kids.
Much appreciated! I gave up on seafile, went back to syncthing.
No problem, I was going to use SyncThing but for my use case, Seafile just came out on top.
If you still have the installation there is a guide on my profile, which I made on how to recover libraries from a corrupted installation etc.
Technically should all still be there as it’s written in block storage so as long as you didn’t physically delete the block it should be A-okay.
Let me know if you do need anything!