Brett brettinternet

## raspberry-pi-4-read-write-speedtest-results.sh
# Micro SD

## read

sudo hdparm -Tt /dev/mmcblk0
/dev/mmcblk0:
 Timing cached reads:   2190 MB in  2.00 seconds = 1095.17 MB/sec
 Timing buffered disk reads: 126 MB in  3.04 seconds =  41.46 MB/sec

## write

## pi-frame.sh
# 1. Setup raspbian lite - no desktop environment or Xorg necessary. Install `fbi`.

# 2. Sync files to Pi:

rsync -ahP --delete ./slides remote-ssh-host:~

# 3. Append to ~/.profile to run on startup

if [[ $(fgconsole 2>/dev/null) == 1 ]]; then
    fbi -d /dev/fb0 -noverbose -a -t 60 -cachemem 0 $HOME/slides/*

## zfs-supported-kernel-version.sh
#!/bin/sh

curl -sS 'https://raw.githubusercontent.com/openzfs/zfs/master/META' | awk '$1 == "Linux-Maximum:" { print $2; exit }'

## compose.yaml
version: "3"
name: utils

x-options: &options
  environment: &environment
    TZ: "${TIMEZONE}"
  restart: unless-stopped
  networks:
    - utils

## Taskfile.yaml
# install go-task
---
version: "3"

vars:
  PROJECT: myproject
  SECRETS_FILE: sops.env
  COMPOSE_FILES:
    sh: find . -type f -iname "compose*.yaml" -exec basename {} ';' | sort -r
  COMPOSE_FILES_ARGS: -f {{ .COMPOSE_FILES | splitLines | join " -f " }}

## oops-zfs.md

      
              1 file
            
          
              0 forks
            
          
              0 comments
            
          
              0 stars
            
          
                brettinternet
                / oops-zfs.md
            
            
              Last active
              December 10, 2023 02:35
            
              
                Recovery from ZFS oops
              
          
    The stages of ZFS grief

Disk passthrough to a VM managing my ZFS array.
$ qm set 100 -scsi1 /dev/disk/by-id/…
$ qm set 100 -scsi2 /dev/disk/by-id/…
$ …

  
## repo_connection_reaper.ex
defmodule MyApp.RepoConnectionReaper do
  @moduledoc """
  Disconnect all connections after a X minutes to reset cache build-up and leaky memory on long-lived connections
  Especially useful for dynamic queries which contribute to this build-up.

  Alternatively, use this Repo option e.g. `Repo.all(query, prepare: :unnamed)` to avoid using named prepared statements
  on queries that are dynamic enough to not cache the prepared statement on the connection.

   See also
   - https://github.com/elixir-ecto/db_connection/issues/99

## map_dev_id.sh
#!/bin/bash

find /dev/disk/by-id/ -type l | xargs -I{} ls -l {} | grep -v -E '[0-9]$' | sort -k11 | cut -d' ' -f9,10,11,12

## slack_cleaner.py
#!/usr/bin/python

from slack_cleaner2 import *

s = SlackCleaner('TOKEN')
channel = "CHANNEL_NAME"

for msg in s.c[channel].msgs(with_replies=True):
    if msg.user == s.myself:
        print(f"DELETED: {msg.text}")

## react-query-derive-maybe-user.tsx
// Nullable user example:

type MaybeUser = UserQuery['user']

/**
 * Possibly null user, such as on an auth page
 * Here's a description on why these overloads are required: https://stackoverflow.com/a/67640634
 */
export function useMaybeUser<UserData = MaybeUser>(
  select: (user: MaybeUser) => UserData
	# Micro SD

	## read

	sudo hdparm -Tt /dev/mmcblk0
	/dev/mmcblk0:
	Timing cached reads: 2190 MB in 2.00 seconds = 1095.17 MB/sec
	Timing buffered disk reads: 126 MB in 3.04 seconds = 41.46 MB/sec

	## write
	# 1. Setup raspbian lite - no desktop environment or Xorg necessary. Install `fbi`.

	# 2. Sync files to Pi:

	rsync -ahP --delete ./slides remote-ssh-host:~

	# 3. Append to ~/.profile to run on startup

	if [[ $(fgconsole 2>/dev/null) == 1 ]]; then
	fbi -d /dev/fb0 -noverbose -a -t 60 -cachemem 0 $HOME/slides/*
	#!/bin/sh

	curl -sS 'https://raw.githubusercontent.com/openzfs/zfs/master/META' \| awk '$1 == "Linux-Maximum:" { print $2; exit }'
	version: "3"
	name: utils

	x-options: &options
	environment: &environment
	TZ: "${TIMEZONE}"
	restart: unless-stopped
	networks:
	- utils
	# install go-task
	---
	version: "3"

	vars:
	PROJECT: myproject
	SECRETS_FILE: sops.env
	COMPOSE_FILES:
	sh: find . -type f -iname "compose*.yaml" -exec basename {} ';' \| sort -r
	COMPOSE_FILES_ARGS: -f {{ .COMPOSE_FILES \| splitLines \| join " -f " }}
	defmodule MyApp.RepoConnectionReaper do
	@moduledoc """
	Disconnect all connections after a X minutes to reset cache build-up and leaky memory on long-lived connections
	Especially useful for dynamic queries which contribute to this build-up.

	Alternatively, use this Repo option e.g. `Repo.all(query, prepare: :unnamed)` to avoid using named prepared statements
	on queries that are dynamic enough to not cache the prepared statement on the connection.

	See also
	- https://github.com/elixir-ecto/db_connection/issues/99
	#!/bin/bash

	find /dev/disk/by-id/ -type l \| xargs -I{} ls -l {} \| grep -v -E '[0-9]$' \| sort -k11 \| cut -d' ' -f9,10,11,12
	#!/usr/bin/python

	from slack_cleaner2 import *

	s = SlackCleaner('TOKEN')
	channel = "CHANNEL_NAME"

	for msg in s.c[channel].msgs(with_replies=True):
	if msg.user == s.myself:
	print(f"DELETED: {msg.text}")
	// Nullable user example:

	type MaybeUser = UserQuery['user']

	/**
	* Possibly null user, such as on an auth page
	* Here's a description on why these overloads are required: https://stackoverflow.com/a/67640634
	*/
	export function useMaybeUser<UserData = MaybeUser>(
	select: (user: MaybeUser) => UserData