FEEDBACK WANTED: Insync 3 headless

Hi David,

Thanks for the kind words. Cool use case by the way! I’ve talked to some of our users about headless lately and they’ve mentioned a similar use case: using headless or Insync remotely to sync/backup their files, generally speaking (wondering exactly how many users have this use case as well though).

Are there any hiccups or frustrations in your current workflow? Anything you find inconvenient?

What would your workflow be without Insync headless? What workaround would you have to do?

1 Like
  1. Why do you need Insync headless?
    Maybe to run a linux server to keep a copy of my Drive files on a Raspberry Pi
  2. What is your workaround without Insync headless? I use Syncthing, is headless but also have a web GUI very well developed.
  3. What features/behavior would make Insync 3 headless better than Insync 1.5 headless? RELIABILITY
  4. Would you prefer to have Insync 3 headless as a separate product or integrated into the GUI? Separate
  5. What platform(s) do you need Insync headless to work? arm Linux in deb packages but also with an apt server
2 Likes

Thanks a lot for this initiative!

2 Likes

Hi Josh–
As I mentioned previously that I would have to fire up a desktop session , either through the console or through vnc for insync to work. Otherwise, my files will not be backed up to the cloud. And yes that is annoying.
I thought of another use case , sort of similar to my own. As you may know , *nix doesn’t need a desktop. This is typical in the server space. So many people use a Linux system as an appliance. For example, insync could sync to this appliance and in-turn the appliance could sync to the cloud for backup and/or act as a local file server. All without a Desktop Environment. That would be pretty cool.

  • Why do you need Insync headless?
    I run a Linux server (no GUI) at home that acts as my file server where I use insync-headless 1.4.9 to keep a local copy of my Google Drive. That is then shared on the network to any of my devices/PCs that need it, instead of keeping multiple copies of my 2TB Google Drive on multiple devices.
  • What is your workaround without Insync headless?
    When I first set this up insync-headless was the best solution I could find, so I don’t know.
  • What features/behavior would make Insync 3 headless better than Insync 1.5 headless?
    insync-headless 1.4.9 has been pretty solid. There was an issue that broke it when I upgraded to 1.5 though. Just general reliability. Perhaps some native email alerting, though that’s easy to handle outside of the program.
  • Would you prefer to have Insync 3 headless as a separate product or integrated into the GUI?
    I would want no dependencies on GUI elements/functionality, so I’m inclined to say separate.
  • What platform(s) do you need Insync headless to work?
    CentOS/RHEL
1 Like

Hi Gregory,

  • Why do you need Insync headless?
    Strictly speaking I don’t need it - I have a working solution, but my usage case is to backup files from my NAS to Google Drive

  • What is your workaround without Insync headless?
    Currently I’m using rclone. I was using the older insync-portable; but after a RAID failure insync decided it wanted to resync 500gb+ of data, so I used rclone instead to avoid unnecessary syncing

  • What features/behavior would make Insync 3 headless better than Insync 1.5 headless?
    As above - better logic on when to resync…

  • Would you prefer to have Insync 3 headless as a separate product or integrated into the GUI?
    Don’t care.

  • What platform(s) do you need Insync headless to work?
    Suitable for Debian Buster on the Pi 4 for me

1 Like

Thanks so much for the feedback everyone! Keep 'em coming :slight_smile:

*** Why do you need Insync headless?**
I have two use cases for the insync headless:

  • Using it on headless NAS server to sync with Gdrive. Ideal for backup and saving space by not copying the same files on multiple machines.
  • Using insync in general. No offense but I think the insync GUI is not useful at least to me. It scales poorly with display resolution and I want to make my file management with my file manager. That’s why I use insync in first place - to browse “cloud” files locally. The only time i need a GUI is mark the files/folder which I don’t want synched in particular machine.
  • External configuration file so instances can be migrated easily

*** What is your workaround without Insync headless?**

  • I am using rclone but insync is more “set it and forget it” than rclone which requires manual sync scheduling.

*** What features/behavior would make Insync 3 headless better than Insync 1.5 headless?**

  • Keep the command line “GUI” from the previous version
  • Improved logging - this is general suggestion not only for the CLI. Keeping the logs in proprietary format is inconvenient.
  • Clear error reporting - write a syslog message when sync fails, send a mail, whatever.
  • Improve progress reporting. When syncing a drive fresh you have no idea what is going on, especially if the drive is big.

*** Would you prefer to have Insync 3 headless as a separate product or integrated into the GUI?**

  • Separate. In the case of headless server it shouldn’t drag unnecessary X11 and so on dependencies.

*** What platform(s) do you need Insync headless to work?**

  • Arch, Ubuntu. I understand if you don’t provide native Arch support and don’t worry sooner that you think it will be in AUR :slightly_smiling_face:
2 Likes
  1. Why do you need Insync headless?
    So that I can have a local copy of my Google drive files on a headless Linux file server. Largely to make sure that I have redundant backups in different places of different types.
    I also use the Linux server for a variety of things like configuring cloud VMs, scripting, etc. from the command line, and want my files on the server backed up (to Google drive).

  2. What is your workaround without Insync headless?
    I have none (I use v1.5.7)

  3. What features/behavior would make Insync 3 headless better than Insync 1.5 headless?
    I just need it to keep an accurate sync of my files

  4. Would you prefer to have Insync 3 headless as a separate product or integrated into the GUI?
    I have no GUI

  5. What platform(s) do you need Insync headless to work?
    Ubuntu

1 Like

Why do you need Insync headless?

  • Used on a openmediavault server; which does not have a GUI.

What is your workaround without Insync headless?

  • No workaround, using v1.5.7

What features/behavior would make Insync 3 headless better than Insync 1.5 headless?

  • Reliable, Reliable, Reliable, Reliable syncing, with no loss of data.
  • Advanced logging for bot normal operation and error conditions.
  • Easily accessible configuration file.

Would you prefer to have Insync 3 headless as a separate product or integrated into the GUI?

  • Separate to avoid the need for unnecessary library includes/dependencies. Also need to be able to install via CLI.

What platform(s) do you need Insync headless to work?

  • Linux - Debian, Ubuntu, Arch & RedHat derivatives (CentOS, Fedora).
1 Like

I use this custom made docker image to run insync on a server in order to backup Google Drive folder: https://github.com/hadim/docker-insync

1 Like
  • Why do you need Insync headless?
    I sync between my main desktop and laptop when traveling, and when doing so I ssh into my main desktop, and move the files I need to my insync directory. Previously this wasn’t an issue because I could just use the command line interface to start insync on my main desktop remotely via SSH. Now I can’t do that and it’s a PITA to have to start up a VNC session just to start insync. Also, as you may or may not remember, there was an issue with insync not starting with Fedora 30. The workaround was to add a statement at system startup to start insync headless and use the command line interface. If you don’t have a headless mode, you lose a potential circumvention for GUI issues. And, this issue wasn’t fixed quickly - which highlights the need for a command line interface if for nothing else, circumvention purposes.
  • What is your workaround without Insync headless?
    The workaround is to use VNC for file transfer. However, there is no solution if for whatever reason the GUI stops working.
  • What features/behavior would make Insync 3 headless better than Insync 1.5 headless?
    I can’t think of any, the commands were fairly comprehensive. You should at least keep functionality parity with 1.5 - why would you have a later release with less functionality?
  • Would you prefer to have Insync 3 headless as a separate product or integrated into the GUI?
    It should be the same product and incorporated into the GUI as it was previously. Again, I don’t understand the strategy of coming out with a later release of something with less functionality. Insync should always have a command line interface.
  • What platform(s) do you need Insync headless to work
    Linux, specifically the current releases of Fedora.
2 Likes

Woot! I’m glad to hear you guys are looking into this! I’ve been using the headless version for quite some time and have come to rely on it!

Why do you need Insync headless?

I use it on a system that doesn’t have a GUI on it so a shell via ssh is my only interface. It’s actually a virtualized server. I have several pieces of automation via bash scripts where I rely on get_status, get_sync_progress, pause_syncing, resume_syncing and so forth. I also use the manage_selective_sync interactively, so the curses-like UI for that is very helpful.

What is your workaround without Insync headless?

Right now, I’m exploring that. I’ve been using v1.4.9 but have had issues with it lately and haven’t been able to solve them.

What features/behavior would make Insync 3 headless better than Insync 1.5 headless?

The same features I mentioned above would be great. If there was better memory management, even better.

Would you prefer to have Insync 3 headless as a separate product or integrated into the GUI?

Doing it separately would probably be better in my scenario so installing the package wouldn’t try and pull in all the GUI dependencies.

What platform(s) do you need Insync headless to work?

Debian would be my first choice as that’s what I use now. If I had to, I could do Ubuntu.

Thanks guys for taking those of us who use(d) the headless version into consideration!

2 Likes

Continuing the discussion from FEEDBACK WANTED: Insync 3 headless:

  1. Why do you need Insync headless?
    Because I can run batch jobs on an unattended machine.

  2. What is your workaround without Insync headless?
    At the moment, I’m using an older version. Works fine, BTW.

  3. What features/behavior would make Insync 3 headless better than Insync 1.5 headless?
    Being supported is more than enough!

  4. Would you prefer to have Insync 3 headless as a separate product or integrated into the GUI?
    Don’t care but being integrated could lead to installation problems… (e.g. libraries compatibility)

  5. What platform(s) do you need Insync headless to work?
    Linux. Maybe useful for MacOS, I don’t know. Useless for windoze.

  • Why do you need Insync headless?
    My use case revolves around being able to have workflows that are downstream of Google Drive files. I have a NAS server (Ubuntu) which I run and I use Google Drive as backup/pipeline solution. It’s a bit difficult to explain but as it stands, I have to have a GUI installed on that machine just for Insync. The problem is that every time I run into an error (what prompted this is that I’m getting an OS Error because one of my drives is full. The GUI fails silently and I had to physically walk into that room, connect a monitor, poke around etc. instead of being able to connect over a terminal like I normally do).

  • What is your workaround without Insync headless?
    Install a distribution with a GUI solely to use Insync.

  • What features/behavior would make Insync 3 headless better than Insync 1.5 headless?
    I didn’t use Insync 1.5 headless, I must have purchased after that was deprecated.

  • Would you prefer to have Insync 3 headless as a separate product or integrated into the GUI?
    It might be nice if it was a separate product that was integrated with the GUI application. E.g. you could run the GUI if you had it on a machine but it still let you alter your settings from the command line. So on the machine I have it on now, I have a GUI so I could configure it that way but if I had an issue or wanted to change settings, I could SSH in. That would be in contrast to installing a separate product that I would have to use one or the other. That said, I would be fine with that approach as well.

  • What platform(s) do you need Insync headless to work?
    Linux (Ubuntu)

1.5 headless is not deprecated and is the one we recommend you use for now

  1. Backup personal gdrive on headless home server.
  2. Rclone. Works well enough, but have encountered incompatibilities between insync and rclone conversion of Google Docs to MS Office format, so it’s better to use the same product for desktop and server sync.
  3. Regular security updates and bug fixes would be a good thing.
  4. If the gui and headless functionality could be integrated, that would be good: but I could live with them remaining separate products.
  5. Linux: Latest Ubuntu LTS, CentOS release, Fedora.
  1. Syncing my Windows Subsystem for *nix Profile between Workstations.
  2. Currently I can do this using the windows Version of it, but Problem is Permissions get f*ed up each time it needs to update something. Don’t sync certain files that get executed automatically, on startup to reset the current Permissions.
  • Why do you need Insync headless?

Because I automate everything in my workstation, including initial setup with Ansible. Plus, I generally prefer setting up stuff with cli, which is extensible for automation.

  • What is your workaround without Insync headless?

Well, none!

  • What features/behavior would make Insync 3 headless better than Insync 1.5 headless?

I haven’t checked 1.5 headless features, but I would need a systemd service launched at statup with --user option.

  • Would you prefer to have Insync 3 headless as a separate product or integrated into the GUI?

Separate, I guess. The client / daemon architecture would fit nicely, with headless being just a second client. You would have an insyncd daemon launched by systemd in user land, an insync-headless to interact with the daemon through cli, and insync-gui or insync the GUI program.

  • What platform(s) do you need Insync headless to work

Mainly Arch and Debian. Perhaps MacOs.

Why do you need Insync headless?

  • I use it (several subscriptions on different accounts) to synchronize files on headless Linux servers to/from G-Drive for backups and automated updates between those Linux Servers (running Insync) and various Windows 10 desktops (running Google Backup & Sync and/or Google Drive File Steaming).
  • ​My ONLY use-case for Insync IS headless.
  • Google’s native products (Backup & Sync, and Drive File Streaming) don’t provide that.
  • Google’s native products (Backup & Sync, and Drive File Streaming) provide all OTHER functionality I really need EXCEPT headless.

What is your workaround without Insync headless?

  • ​NONE, using very old version.​

What features/behavior would make Insync 3 headless better than Insync 1.5 headless?

  • Just keep it updated, maintained, supported on latest operating systems, file-stream and any other target support, etc.

Would you prefer to have Insync 3 headless as a separate product or integrated into the GUI?

  • ​Definitely no requirement for any GUI​ - i.e. remain TRULY “headless” where the only access is SSH and no GUI libraries, etc.

What platform(s) do you need Insync headless to work?

  • ​Definitely “headless” for Ubuntu LTS - 16.x, 18.x and soon-to-be 20.x​
  • Likewise for Debian, SteamOS, etc.
  • Identical “headless” for Window Server 2016/2019 would also be helpful.