Web 1.5.0 on Windows10. Starts as a service, but i can't access the WebUI

I had setup Duplicacy on this laptop months ago (and have a log to show it), but i realised recently it hadn’t been running it’s backups. So i jump on the laptop and can’t access the WebUI: “unable to connect”.

I fired up the App, and i could access the GUI by clicking on it and “Open” and it asked me for a new password. I enter the same as i used last time and i find that all my previous settings were gone :frowning: The app icon is grey too btw.

I reboot, and i see it is running as a service. But same thing, can’t get in unless i use the App. Although strangely i can access 127.0.0.1:3875/setting (no other sections) before loading the app.

I’m content to input all my settings again, but it doesn’t look like it starts with the laptop, or will run it’s jobs - so i don’t want to go through all that before i can access the GUI.

I tried two browsers - Edge and Firefox in case an addon might have blocked it.

I even uninstalled it as administrator and re-installed it as administrator. Reboot - service is running. Go to browser and i can’t access 127.0.0.1:3875

Any ideas? Thanks.

In the c:\ProgramData there will be duplicacy folder with logs. check duplicacy_web.log, to confirm whether it succeeds listening on a port on which one.

Thanks Saspus.

Here’s the first failure:

2021/07/08 03:16:30 Set current working directory to C:\ProgramData/.duplicacy-web/repositories/localhost/0
2021/07/08 07:14:11 Failed to parse the configuration file: invalid character '\x00' looking for beginning of value

And it’s what continues now.

I haven’t changed any settings in the backups that might have done anything that i can recall.

Filesystem corruption?

Do this: in the elevated command prompt run chkdsk c: /f, confirm with y and reboot. During next boot windows will run disk check. It will then attempt to fix discovered inconsistencies and if successful — will reboot again. I see windows volumes getting corrupted often that can lead to many crazy issues.

Then delete everything under C:\ProgramData/.duplicacy-web/repositories. That’s is a temporary location. Nothing there is important.

Then check duplicacy_web.log produced by web gui for the listening port, and review settings.json and duplicacy.json visually — those are text files and should look like nonempty text files with key-value json structure. They should be in the same folder as .duplicacy_web, but I’m not sure, I don’t have windows to check on.

1 Like

Thanks again @saspus.

I did all the things and it was still busted. But I probably should have done the last one first as i can see that duplicacy.json is empty - well, not quite - it is 6kb and appears to have lots of spaces (it’s selectable) but zero text. Looks like the culprit by your description.

The repository directory has all my settings seemingly intact, so i’ve put that back.

EDIT:
I deleted the apparently corrupt duplicacy.json and i can now load the web-ui. It looks too difficult for me to input the settings into that file unless i have that file (it’s not in my backups).

How so? All you need is credentials for storage and storage password. Add the storage and then create new backup schedule. Takes 2 minutes and a few clicks. Much faster than messing with restoring duplicacy.json and restarting the service, etc. FWIW I don’t even think about backing up duplicacy configuration myself. Just save the credentials in the password manager and/or keychain

I don’t know what a working one looks like, i don’t know the syntax. I don’t know what “encryption data” means. There’s a lot more settings in there than just credentials and password.

{
    "storages": null,
    "computers": [
        {
            "name": "localhost",
            "repositories": null
        }
    ],
    "schedules": null,
    "key_iterations": 16384,
    "encryption_data": "REMOVED FOR SECURITY(?)",
    "cli_version": "",
    "administration_token": "",
    "administration_token_expiration": 86400,
    "license_id": ""
}

It’s ok, i’ll just do it the long way.

In all fairness you are not supposed to know how it looks like and what entry mean. Those are implementation details. Duplicacy manages that data on its own.

Actually, it’s the short way. It takes just a few clicks to input license, add storage, and configure schedule, and it’s way faster than trying to reverse engineer and emulate the configuration file and then hope that you did it right and not miss anything that could bite later.

The advice to look at that file was solely to check for obvious corruption, because it’s easy to see as the file is supposed to only have printable characters and human readabale data.

Thanks for your help.

It would be easier to update the json if i knew what i was doing :slight_smile:

I’ve set it all up again, and it’s all looking ok now… and i’ve included the .duplicacy-web directory in my backups :smiley: It’s only a few mb.

Not sure if that’s a great idea, due to the cache. What do you think?

One day i should set up the notification service that lets me know if it doesn’t run at all. It’s been on my to-do list for… well, since i bought Duplicacy. I’ve had a quick look and then i get distracted. Like right now…

I would simply copy duplicacy.json and settings.json to some other folder, that is being backed up.

If I remember correctly duplicacy ignores its own cache folders, so technically it should take care of itself :slight_smile:

This is super easy to do – do it now :). Create project at https://healthchecks.io and get update URL. Add it to the Backup properties (not Shedules, specifically backup)

This topic was automatically closed 10 days after the last reply. New replies are no longer allowed.