Posts tagged: blog

Shotput

2025-07-16T06:18:00Z

This is from I built a simple app for adding images to my blog

Overview #

Shotput is a simple webapp backed by a git repo that allows you to paste a screenshot from your clipboard, it gets put into the repo and pushed, and you get a link back to using that image via statically’s CDN in your website.

Example #

pasting an image let’s you copy the markdown link for easy pasting

![20250607130609_dadd33eb.png](https://cdn.statically.io/gh/pypeaday/images.pype.dev/main/blog-media/20250607130609_dadd33eb.png)

Concerns #

Currently this is deployed with an ssh key mounted into the container that I put into a zfs dataset manually. Per 2025-07-16 Notes I’m struggling with the key sharing model…

💭 Tyblog | systemd has been a complete, utter, unmitigated success

2025-07-11T11:00:54Z

Here’s my thought on Tyblog | systemd has been a complete, utter, unmitigated success

I found this post on systemd this morning, and although I’ve been a Linux user for around 6 years now, there’s some levels of it that are pretty foreign to me. One of the things I never got was the drama around systemd - but that’s because I wasn’t around in the beforefore times, I don’t use cron THAT MUCH, and I’ve only had to setup a few systemd unit files - not enough to generate a real opinion on it. But lately I’m looking into more automation at home - and some stuff is fun to put into an overkill of a platform for a homelab, like temporal, but some stuff should probably be run in more hardened ways (backups for example) - so getting a little download this morning on systemd from a clearly seasoned Linux user is a nice addition to my hot cup of coffee.

Note

This is one of [[ my-thoughts ]]. I picked this up from Waylon Walker(https://thoughts.waylonwalker.com). It’s a short note that I make about someone else’s content online. Learn more about the process Thoughts

💭 { TechDufus } | Blog

2025-07-02T11:09:01Z

Here’s my thought on { TechDufus } | Blog

Just came across TechDufus today… love his blog style… it’s beautiful… definitely not gonna potentially rip this off…

I like the estimated reading time
the tree/ls like view of the posts is super cool

Note

💭 A Brief History & Ethos of the Digital Garden

2025-06-17T11:56:10Z

Here’s my thought on A Brief History & Ethos of the Digital Garden

I am interested in the topic of “digital gardening”, and I didn’t realize how long the term has been around. As a relatively new member of the blogging community I’d like to formulate my own understanding and continue to grow in practicing tending to what I put out their in the digi-space.

Note

💭 Visualizing My Blog’s Internal Links - Jim Nielsen’s Blog

2025-06-16T11:26:29Z

Here’s my thought on Visualizing My Blog’s Internal Links - Jim Nielsen’s Blog

Someday I will build a thing to see the relationship in my posts and notes. Right now I’m working on writing consistently, improving my tagging, using wikilinks, and just generally thinking about what I want my website to contain - and a mindmap like this would be just awesome

Note

💭 slash pages

2025-06-12T11:41:43Z

Here’s my thought on slash pages

Waylon’s mentioned slash pages before, but as I work on building my blog again and writing more regularly thanks to “thoughts” etc. I’d like to start putting together some meaningful slash pages for myself… I’m not sure what to put out there yet but this thought can serve as a todo for me

Note

Recovering OPNSense

2024-11-06T00:00:00Z

I woke up to faulty internet and after some troubleshooting it turns out the root zfs dataset that OPNSense boots from got corrupted…

PRO-TIP - Auto backup your OPNSense config to Google Drive, git, or nextcloud… But if you won’t then at least back up your OPNSense config somewhere everytime you update it.

It’s too much to recount every issue, so here’s a bullet list what worked.

On a fresh drive install OPNSense
Plug in the old drive through a USB enclosure - now I’m not sure what would happen if you plugged it in along with the new drive and then booted up. Because both drives will have a zfs pool zroot and the boot dataset is automounted at /zroot/ROOT/default. My old zroot pool was SUSPENDED so it didn’t automount
Because the old zoot/ROOT/default was corrupted I did this to mount it RO: zpool import -d <path to zfs partition - /dev/stuff> -N zroot zrootrecovery

-d is the zfs flag to import the pool by disk id, -N it to not mount any of the datasets (we need to change mountpoints) and the zroot zrootrecovery imports the zroot pool with a new name

Change the mountpoints for all the zrootrecovery datasets to somewhere like /mnt/zrootrecovery
Depending on the mount point you set you’ll find a config directory around /mnt/zrootrecovery/ROOT/default/config - copy the file you want to another machine via scp or whatever
Go to OPNSense webui and recover from that config!

All in all this process took me around 8 hours but I did run into about ever issue under the sun (several bad disks in the mix, a laptop that wouldn’t live boot into a BSD system, etc.)

Wish-List-With-Fastapi

2022-05-06T00:00:00Z

Amazon has crossed the line with me just one too many times now so we are looking to drop them like every other Big Tech provider….

However, one key feature of Amazon that has been so useful for us is Lists… We can just maintain a list for each of us and then family members can login anytime and check it out… This really alleviates any last minute gift idea stress right before a birthday or something.

So I need a nice gift list service but I don’t want to be locked into one company (like a Target registry or something) and I’d like to host it myself

The internets had a few options but nothing looked/felt like I wanted to I decided to build my own.

The Frontend

I have no idea how to do front end so stay tuned

The Backend

FastAPI for the win on this one… I followed a few examples online and what I was able to build in just a few minutes is pretty impressive thanks to the design of FastAPI.

Some key features are:

Auto doc generation
Required typing (which makes #1 possible)
Built-in api testing in the browser
Easy integration with sqlalchemy
Development time so short you won’t be done with your coffee before having something up and running!

Database #

Starting with a simple database.py we can create a sqlalchemy session with a base model with about 7 lines of code…


from sqlalchemy import create_engine
from sqlalchemy.ext.declarative import declarative_base
from sqlalchemy.orm import sessionmaker


SQLALCHEMY_DATABASE_URL = "sqlite:///wishes.sqlite3"
engine = create_engine(SQLALCHEMY_DATABASE_URL)
SessionLocal = sessionmaker(autocommit=False, autoflush=False, bind=engine)
Base = declarative_base()

Model #

For my wish list I needed just a simple table:

id	person	item	link	purchased	purchased_by	date_added
1	pypeaday	A sweet item	www.mystore.store	False	dad	2022-05-05 21:55:09
2	pypeaday	A bitter item	www.bitterstore.com	True	Mrs. pypeaday	2022-05-06 06:55:54

The table is simple enough… A unique key, the person who the wish belongs to, the item (or wish), a link to the item, whether it’s been purchased or not and by whom, and the date it was added.

To make this model with sqlalchemy we can make a model.py like so:

from database import Base
from sqlalchemy.schema import Column
from sqlalchemy.types import Boolean, Integer, String, Text


class Wishes(Base):
    __tablename__ = "Wishes"
    id = Column(Integer, primary_key=True, index=True)
    person = Column(String(20))
    item = Column(Text())
    link = Column(Text())
    purchased = Column(Boolean())
    purchased_by = Column(String(90))
    date_added = Column(String(15))

Schema #

One of the best things about FastAPI is trivial integration with pydantic. We can define a schema to ensure any data posted is not missing anything!

Make a schema.py with the following:

from pydantic import BaseModel
import time
from typing import Optional


class wish_schema(BaseModel):

    person: str
    item: str
    link: str
    purchased: bool = False
    purchased_by: Optional[str] = None
    date_added: Optional[str] = time.strftime("%Y-%m-%d %H:%M:%S", time.localtime())

    class Config:
        orm_mode = True


class patch_schema(BaseModel):

    purchased: bool
    purchased_by: Optional[str] = None

    class Config:
        orm_mode = True

I have 2 schemas - one for a wish which you’ll see down below is used to validate any post requests.

To simplify things for me I made another schema, patch_schema which I use for the route that updates the table (ie. marking an existing wish as purchased)

Session #

One of the last things we need is a Session

So make a session.py…

from database import SessionLocal, engine
import model

model.Base.metadata.create_all(bind=engine)


def create_get_session():
    try:
        db = SessionLocal()
        yield db
    finally:
        db.close()

Our routes will depend on this create_get_session function that will yield a db object through which we’ll udpate our database

Ok just do it already!

So our main.py will have a few routes in it…

What do we want to support?

Getting all wishes
Getting a specific wish
Updating a specific wish
Deleting a wish

I think the script is fairly self explanatory but here’s a few notes…

We decorate each function with @app.<method> and define response_model as well as status_code
The functions are defined with async (this was my first exposure to this so I can’t go in depth on it yet)
The functions all take a db which is from session.py and that db depends on the create_get_session function
If the db is being updtes then we type the object used for the update with the appropriate schema (either wish_schema or patch_schema)

From there we’re in true python-land where you can basically guess the methods on db and you’d probably be right… (like query, upddate, delete etc.)

from fastapi import FastAPI, Depends, HTTPException
from sqlalchemy.orm import Session
from typing import List
from model import Wishes
from schema import wish_schema, patch_schema
from session import create_get_session

app = FastAPI()


@app.get("/")
def read_root():
    return {"message": "server is up!"}


@app.get("/wishes", response_model=List[wish_schema], status_code=200)
async def read_wishes(db: Session = Depends(create_get_session)):
    wishes = db.query(Wishes).all()
    return wishes


@app.post("/wishes", response_model=wish_schema, status_code=201)
async def add_wish(wish: wish_schema, db: Session = Depends(create_get_session)):
    new_wish = Wishes(
        person=wish.person,
        item=wish.item,
        link=wish.link,
        purchased=wish.purchased,
        purchased_by=wish.purchased_by,
        date_added=wish.date_added,
    )
    db.add(new_wish)
    db.commit()

    return new_wish


@app.get("/wishes/{id}", response_model=wish_schema, status_code=200)
async def get_wish(id: int, db: Session = Depends(create_get_session)):
    wish = db.query(Wishes).get(id)
    return wish


@app.patch("/wishes/{id}", response_model=wish_schema, status_code=200)
async def update_wish(
    id: int, patch: patch_schema, db: Session = Depends(create_get_session)
):
    db_wish = db.query(Wishes).get(id)
    db_wish.purchased = patch.purchased
    db_wish.purchased_by = patch.purchased_by
    db.commit()
    db.refresh(db_wish)

    return db_wish


@app.delete("/wishes/{id}", status_code=200)
async def delete_wish(id: int, db: Session = Depends(create_get_session)):
    db_wish = db.query(Wishes).get(id)
    if not db_wish:
        raise HTTPException(status_code="404", detail="Wish id does not exist")

    db.delete(db_wish)
    db.commit()

    return None

My Code

You can find my repo here.

I’ll plan to update and maintain for as long as I use it

Home-Server-Refactor

2022-04-10T00:00:00Z

My current homelab setup is not great but it works…

Proxmox on PowerEdge R610

I boot off an SD card and have 1 SSD and 5 HDDs configured as a JBOD array using a Dell H700 SAS controller. I cannot boot from a disk using this controller and I can’t get the firmware configured in a way to allow it. So I have 1 SSD as a ZFS array that I’ve been putting my VM images on, and the 5 HDDs are passed through to a TrueNAS VM where I handle all the ZFS stuff there… kind of meta because I then attached those drives to Proxmox as a CIFS share.

TrueNAS on dedicated box

I have an on-prem backup that is just an old desktop running TrueNAS I regularly backup the 5 disk RAIDZ2 array from my Proxmox host (managed by a TrueNAS VM) to this backup box

Currently there is nothing else running on this machine since it’s my “backup”

Jellyfin

I was HWA for Jellyfin, but hardware passthrough on the R610 is finicky or broken so Jellyfin is running on an Ubuntu host.

I could put UBuntu on the R610 and give up “true virtualization”. Then I’d manage the SMB share myself. If I do that then I would get rid of “users” I think, ie. basically forgo least-priviledges since I’m not sure how hard that is to manage.

On the other hand, direct access to the smb config might make it easier?

I have the media array on Jellyfin box setup as NFS which was really easy with ZFS… I think SMB would be just as easy.

Plan of attack…

Move all vm disks to individual datasets on the NAS
Backup docker data… not sure how well this will work, maybe just start over?
Clean up Ansible playbooks on the user side of things - stick with neville vs just using my own name?
Install Ubuntu 20 or 22 on a 2.5" drive that I’ll toss in this SSD enclosure (or a usb thumb stick?)
Re-deploy everything with ansible-playbook and configure…

Configuration… #

THE FREAKING NAS -> just import zfs array and configure SMB?

1.~~ Nextcloud users and connections.. might be able to just copy the data folder? not sure about the database… try spinning it up in the sandbox vm and see if stuff is there ~~ 2.~~ *arr suite, media profiles and connections to transmission… nothing major~~ 3. transmission - should be deploy and go 4. ombi and jackett should also just work after some config again 5. ~~traefik should just work~~ 6. ~~try to bring up pi-hole from the vm that’s already running~~ 7. ~~heimdall will hopefully just be copying the data folder from the existind docker one’~~ 8.~~ booksonic can be reconfigured easily~~ 9. ~~portainer… hopefully just copying data folder over?~~ 10. ~~littlelink~~, small-group-notes, and blog (at home) will need manually re-deployed once Ubuntu is installed bare-metal

BIG BIG BIG TODOS #

Sanoid/syncoid! Get snapshots going and backups configured with on prem TrueNAS
Wireguard setup on DA.
network share on printer for paperless ~~4. update peperless in ansible-nas~~
~~Just deploy paperless manually… monitor/manage with portainer~~
booksonic not seeing audiobooks/podcasts
need a smb user to map nas/documents to the printer for paperless
wireguard setup now on kps phone, desktop, server (and backup truenas?), and dad’s pi
~~verify lan services work~~
~~Tdar so Jellyfin can work better~~

Snapshot business might be cause of all the docker containers and docker using ZFS backend… take everything down and try removing

file browser - currently I just one-clicked in portainer, I want to make a stack with my own config file which I’ll rip from techno tip and then add my traefik lables too

Forget filebrowser - going to just use Nextcloud for how it’s supposed to be used. 3. Need to organize those files in nextcloud

CHECK THIS -> ran as ‘cp’ utility in tmux window, no progress bars or anything. it’s in ansible-nas session -> window 3 #

Olivet bible stuff going to /tmp/olivet/ -> will move this to nextcloud, ideally by the app via appimage so that the db updates and I don’t have to run that occ script I wnat to organize “home” still in nextcloud

setup Sanoid

clean up bitwarden learn nextcloud sharing -> maybe just give a link to grandma? rest of todos -> document db and sanoid + zfs.rent

Check on mom’s will do media thing for church - split vocals on mp3/4

permission-data playbook changes everything to ansible-nas:ansible-nas but then samba task will re-permission some stuff to root:users… this looks fine I had to add group to the samba config in my playbook to get user auth to work with samba This isn’t fully working… it works from cli but my python process can’t write to a folder in dump after 777…. need to learn more? So I can make a file after adding the ansible-nas group to config, but I still cannot make a directory on the smb mount…

ADDING inherit permission = yes under [global] in the smb.conf worked!

still not working from printer… I think what I want is to setup 2 scan options - single docs right to paperless, or combined scans to dump, then manually split and send to paperless