3 ms·

I'm about 2 hours into setting up an Mobilizon instance for outdoor socially distant activities, but running into problems [1] and at the end of the day federat

I'm about 2 hours into setting up an Mobilizon instance for outdoor socially distant activities, but running into problems [1] and at the end of the day federation isn't important to me and I'd prefer the learning opportunity and flexibility of building my own [2]. Curious if any HNers in the PNW region might use such a service to pod up for outdoor socially distant activities (hiking, wilderness cleanup, kayaking, perhaps camping/backpacking/biking, etc). I moved here just pre-COVID, and quarantine, my introverted nature, lack of FB, and tendency to enjoy things that don't involve lots of people being around has left me without as much of a do-things-with group as I might like. [1] I'm 2+ hours into provisioning/setup, and unfortunately the yarn install keeps hanging/crashing on the $5 digital ocean droplet I provisioned for this, and at the end of the day I'd prefer to have full knowledge and control over the stack. Additionally, for me federation isn't important, all I want is an open architecture that respects privacy. [2] To keep things simple for myself and users, it wouldn't be federated, instead there would be a single domain I manage that hosts the events. The architecture would of course be open source, and my goal would be to make the entire project be available as some sort of { docker image / kubernetes cluster / DO App / heroku thing / something } such that anyone can easily launch their own instance without needing to dedicate 2+ hours to the process, and more importantly any user can find activities to join in on without needing to set up an account/provide email/etc.

prophesi 6y ago
Once the docker installation is documented, it should take considerably less time to self-host. This is for `yarn build`, but perhaps it applies to `yarn install` as well: there's a note in the documentation that says compiling the front-end requires 2gb of RAM. You'll need to set it to use only 1gb RAM to accommodate the $5 droplet's specs via `NODE_BUILD_MEMORY=1024 yarn build` As for the technology & stack, most of the code is from Pleroma, essentially a lightweight Elixir/Phoenix version of Mastodon, with Vue for the frontend. If there's any part of the architecture that's closed, the project would highly appreciate raising an issue, as it wishes to be a fully libre project.
ivarv 6y ago
Where in the PNW are you located? Given the border closure it's no longer feasible to organize international meetups.
bhattisatish 6y ago
Well, this could be because you are running out of memory, which results in the yarn build process getting killed. The simplest solution is to enable "swap" in your instance by using the following commands: dd if=/dev/zero of=/swapfile bs=1M count=1024 sudo chown root:root /swapfile chmod 600 /swapfile mkswap /swapfile swapon /swapfile add /etc/fstab /swapfile swap swap defaults 0 0 #echo 0 > /proc/sys/vm/swappiness sudo sysctl vm.swappiness=10 add to /etc/sysctl.conf vm.swappiness=10 All I am doing is creating a swap file of 1GB and enabling it with a not so aggressive swappinness on the kernel.

prophesi 6y ago

Once the docker installation is documented, it should take considerably less time to self-host. This is for `yarn build`, but perhaps it applies to `yarn install` as well: there's a note in the documentation that says compiling the front-end requires 2gb of RAM. You'll need to set it to use only 1gb RAM to accommodate the $5 droplet's specs via `NODE_BUILD_MEMORY=1024 yarn build` As for the technology & stack, most of the code is from Pleroma, essentially a lightweight Elixir/Phoenix version of Mastodon, with Vue for the frontend. If there's any part of the architecture that's closed, the project would highly appreciate raising an issue, as it wishes to be a fully libre project.

ivarv 6y ago

Where in the PNW are you located? Given the border closure it's no longer feasible to organize international meetups.

bhattisatish 6y ago

Well, this could be because you are running out of memory, which results in the yarn build process getting killed. The simplest solution is to enable "swap" in your instance by using the following commands: dd if=/dev/zero of=/swapfile bs=1M count=1024 sudo chown root:root /swapfile chmod 600 /swapfile mkswap /swapfile swapon /swapfile add /etc/fstab /swapfile swap swap defaults 0 0 #echo 0 > /proc/sys/vm/swappiness sudo sysctl vm.swappiness=10 add to /etc/sysctl.conf vm.swappiness=10 All I am doing is creating a swap file of 1GB and enabling it with a not so aggressive swappinness on the kernel.