I don’t know how this could happen, but even with a new and empty duplicacy.json file, it is not hard to recreate the same storages and backups. If you set the storage directories and backup ids correctly, all the backups should go to the same locations and not upload existing files again.
YES, it should be documented in the release notes that (at least if running on a windows machine)
- the installation path changed
- the former version has to be manually removed
- if now running as a windows service the configuration files need to be moved (stopping the new service, moving the files, starting the service)
just figured this out and fixed it, but that is learning the quite hard way…
already had a smoother user experience…
i moved my settings over and all seems to be working but I dont trust the backup anymore. The first backup job after moving the settings over it uploaded like 39 or 40 gigs of files that were already uploaded… This is messed up.
This shouldn’t happen.
Are the backup ids the same - you didn’t try to rename them?
What does your log of the backup say? Especially the BACKUP_STATS lines at the bottom.
Does your old or new .duplicacy-web folder still sit inside a repository? How big is this folder?
Run a check job, and if you don’t mind, post the log here and we may be able to figure out why those files were uploaded again.
My Install of the New version went well. i just chose the “Single User Mode” and it Installed over the top of my old one and has worked on 3 different Servers.
gchen,
There’s over 10,000 entries with filenames uploaded. Can I send it do you personally through private message or email?
What I wanted is the log from the check job, not the backup job. Yes, please send it through private message.
so an update. My settings were so screwed up, I ended up nuking everything, clearing out backblaze, uninstalling duplicacy and starting from scratch. I just would never trust my backups with whatever happened.
Its a shamed this happened, not sure what caused it was due to me or due to the installer.
Now i have to wait 2 weeks for my NAS to be backed up again at the upload speed I have.
My guess is that you probably used a different backup id with the new settings so all files had to be rescanned (in which case the upload speed would still be much faster than a completely new backup since most chunks already existed in the storage already). If there were a log from the check job we would be able to to diagnose what actually happened.
1 Like
I still have the logs. I’ll PM them to you.
Your log shows that revision 79 contained 0 files – this can happen if the directory to be backed up is on a network share not mounted correctly due to temporary network errors. As a result, revision 80 had to start from scratch (the equivalent of a new backup id) but as you can see from the log only 14GB out of 1.6T was re-uploaded. If there were revision 81 it would have acted normally as it would use revision 80 as the basis (for determining which files are new)
Did you try the -threads option to speed up the backup?
the files shouldnt have moved or been unavailable. It was a locally mounted hard drive. Weird. And why would it only upload 14 gb out of the 1.7 tb?
Yeah, the -threads switch helped. My upload speed is saturated…
I know I might have been okay, but I was getting some other weirdness and I just decided to nuke and pave. i.e. I think I had the service and executable running at the same time accidentally, etc.
Because other chunks already existed in the storage.
but that’s my point. Those chunks should have existed as well (the ones uploaded) when I look through the file list. Those are files that were already uploaded before.
This is because of how Duplicacy packs and splits files into chunks. For incremental backups, only new or modified files are packed and then split into chunks. For a new backup that doesn’t have a previous revision, all files are packed and split, and this will cause some new chunks to be created. In your case only less than 1% of chunks were new chunks which is in fact a very good number and shows how effective the splitting algorithm is.
For more details please take a look at Chunk size details
4 Likes
Just wanted to chime in that the exact same issue happened to me. I had been running 1.0.0 with no issues for months. I saw that there was a new version and installed it, using default settings. Everything seemed to work without any problems so I thought all was good. It seems a few days later, I rebooted the system for an unrelated reason and the next time I checked Duplicacy, it was acting like it was a new install. After some information from this tread, I confirmed that I now had Duplicacy database both in my %programdata% folder and my user folder. When I had rebooted, the 1.1.0 version started as a service, looking for database in the %programdata% folder, instead of my user folder as it had been previously.
To get back up and running again, I completely uninstalled everything I could find with Duplicacy (btw, the WIndows service didn’t remove itself during the uninstall) and then installed the 1.1.0 application again choosing the option that did not involve running as a service (can’t remember the exact option but it was like run under user or something). Afterwards, I started Duplicacy and everything was back to normal and running fine.
Basically, the issue comes down to having different database storage locations based on whether or not you install the application as a service or to run under a user and when coming from 1.0.0 to 1.1.0, the installer doesn’t know what you had already set up previously.
If this is so, @gchen, wouldn’t it be a good idea for duplicacy to check for settings in all possible places and inform the user if it finds something in a different location than where the user currently intends to install?
Kind of dumb question, but I dont want to risk messing anything up after a 2 week long re-upload of my entire NAS.
Somehow Duplicacy is set to run as the logged in user, how do I switch it to run as a service? If I do so will it lose settings again?
This is what I did on my brother’s PC…
Right-click the Duplicacy icon tray and Quit. Then uninstall Duplicacy Web Edition. You won’t lose settings.
To re-install Duplicacy Web Edition v1.1, make sure to right-click the installer and ‘Run as administrator’, choose ‘Install for anyone using this computer’ and, right as the end, choose to install as a service.
Here you may have to stop the running Duplicacy service before moving your old settings across (open Task Manager > Services tab, or open Services -> stop the service).
Move, or better yet, copy the C:\Users\<user>\.duplicacy-web directory into C:\ProgramData\.duplicacy-web - overwriting the files there. (You may have to show hidden files or you can just type %programdata% into the Windows Explorer address bar to get there.)
Start the Duplicacy service. There’ll be no icon tray, but you can open http://127.0.0.1:3875/ in a browser. You may have to enter your encryption password.
Thinking about it, you can probably do the steps in reverse - i.e. copy the .duplicacy-web to C:\ProgramData before re-installing. It’d likely pick up the settings without you having to faff around with stopping/starting the service.
2 Likes