mirror of
https://github.com/Brandon-Rozek/website.git
synced 2024-09-18 13:55:12 -04:00
Added medium syndication metadata
This commit is contained in:
parent
79329aae66
commit
387dd491b1
30 changed files with 175 additions and 134 deletions
|
|
@ -1,10 +1,12 @@
|
|||
---
|
||||
title: "Code in LaTex Beamer"
|
||||
date: 2022-01-26T23:11:37-05:00
|
||||
date: 2022-01-26 23:11:37-05:00
|
||||
draft: false
|
||||
tags: ["LaTex"]
|
||||
math: false
|
||||
medium_enabled: true
|
||||
medium_post_id: d91f704bb53a
|
||||
tags:
|
||||
- LaTex
|
||||
title: Code in LaTex Beamer
|
||||
---
|
||||
|
||||
I commonly use the [`listings` package](/blog/latexcode/) to showcase code in my LaTex documents. I tried doing the same in my Beamer slidedecks and I ran into an issue where the LaTex source code failed to compile. After digging around, I figured out its because every slide or frame that includes code (or any verbatim environment) needs to be marked as `fragile`. A minimal example is presented below:
|
||||
|
|
@ -32,5 +34,4 @@ I commonly use the [`listings` package](/blog/latexcode/) to showcase code in my
|
|||
\end{lstlisting}
|
||||
\end{frame}
|
||||
\end{document}
|
||||
```
|
||||
|
||||
```
|
||||
|
|
@ -1,10 +1,12 @@
|
|||
---
|
||||
title: "Code alongside Output in LaTex"
|
||||
date: 2022-01-27T09:15:10-05:00
|
||||
date: 2022-01-27 09:15:10-05:00
|
||||
draft: false
|
||||
tags: ["LaTex"]
|
||||
math: false
|
||||
medium_enabled: true
|
||||
medium_post_id: c4d061953056
|
||||
tags:
|
||||
- LaTex
|
||||
title: Code alongside Output in LaTex
|
||||
---
|
||||
|
||||
Working on a tool paper, I had the desire to share code alongside its output. With a combination of the `listings`, `caption`, and `color` packages, we can get a result that's nice looking.
|
||||
|
|
@ -147,4 +149,4 @@ print(factorial(5))
|
|||
|
||||
\end{document}
|
||||
|
||||
```
|
||||
```
|
||||
|
|
@ -1,10 +1,12 @@
|
|||
---
|
||||
title: "Dafny v3.3 Show Countermodel"
|
||||
date: 2022-02-04T19:43:12-05:00
|
||||
date: 2022-02-04 19:43:12-05:00
|
||||
draft: false
|
||||
tags: ["Formal Methods"]
|
||||
math: false
|
||||
medium_enabled: true
|
||||
medium_post_id: b1976dcfde09
|
||||
tags:
|
||||
- Formal Methods
|
||||
title: Dafny v3.3 Show Countermodel
|
||||
---
|
||||
|
||||
*Warning: `extractcounterexample` is a new flag in Dafny and the command to extract it will likely change*
|
||||
|
|
@ -102,5 +104,4 @@ Running Dafny again will produce:
|
|||
Dafny program verifier finished with 2 verified, 0 errors
|
||||
Compiled assembly into example3.dll
|
||||
Program compiled successfully
|
||||
```
|
||||
|
||||
```
|
||||
|
|
@ -1,10 +1,12 @@
|
|||
---
|
||||
title: "Reasoning through Loops in Dafny"
|
||||
date: 2022-02-05T00:22:58-05:00
|
||||
date: 2022-02-05 00:22:58-05:00
|
||||
draft: false
|
||||
tags: ["Formal Methods"]
|
||||
math: true
|
||||
medium_enabled: true
|
||||
medium_post_id: c0e16f8cd368
|
||||
tags:
|
||||
- Formal Methods
|
||||
title: Reasoning through Loops in Dafny
|
||||
---
|
||||
|
||||
Dafny treats loops like a black box. It could be annoying the first time you experience this and have no clue why the code is not verifying properly. There are two properties that Dafny needs you to prove: partial correctness and termination. Together these form *total correctness*.
|
||||
|
|
@ -111,7 +113,4 @@ method fact(x: int) returns (y: int)
|
|||
y := y * z;
|
||||
}
|
||||
}
|
||||
```
|
||||
|
||||
|
||||
|
||||
```
|
||||
|
|
@ -1,10 +1,12 @@
|
|||
---
|
||||
title: "Docker Secrets"
|
||||
date: 2022-02-04T23:59:13-05:00
|
||||
date: 2022-02-04 23:59:13-05:00
|
||||
draft: false
|
||||
tags: ["Containers"]
|
||||
math: false
|
||||
medium_enabled: true
|
||||
medium_post_id: 58cc9e00b1a7
|
||||
tags:
|
||||
- Containers
|
||||
title: Docker Secrets
|
||||
---
|
||||
|
||||
I try to keep secrets such as passwords and keys out in their own separate files so that I can `.gitignore` them and commit the rest of my configuration. With `docker-compose` we can do that with the `env_file` field. Here is an example with a postgres configuration:
|
||||
|
|
@ -26,5 +28,4 @@ Then in `Volumes/database/docker.env` I can have a file with the secrets as key-
|
|||
POSTGRES_USER=user
|
||||
POSTGRES_PASSWORD=389ed93045c84cc0828c4310e6ef76ce
|
||||
POSTGRES_DB=database
|
||||
```
|
||||
|
||||
```
|
||||
|
|
@ -1,10 +1,12 @@
|
|||
---
|
||||
title: "Finding Cool People on Mastodon"
|
||||
date: 2022-05-15T20:39:35-04:00
|
||||
date: 2022-05-15 20:39:35-04:00
|
||||
draft: false
|
||||
tags: ["Mastodon"]
|
||||
math: false
|
||||
medium_enabled: true
|
||||
medium_post_id: 50c97fc8aa3c
|
||||
tags:
|
||||
- Mastodon
|
||||
title: Finding Cool People on Mastodon
|
||||
---
|
||||
|
||||
Mastodon is a cool federated micro-blogging platform that can serve as an alternative to Twitter. One of its primary features is that it shows what they call toots from only people you follow in reverse-chronological order. Nowadays, large tech companies employ recommendation engines to surface content from users (you don't even need to follow) in a non-chronological order driven by the number of interactions users similar to you had with the content. The lack of recommendation engine in Mastodon can be a blessing as low-quality emotional content tends to drive more interactions. However, this can also be a curse as cool new people don't naturally surface in your feed. The following are the techniques I employ to consistently seek out cool new people:
|
||||
|
|
@ -85,4 +87,4 @@ The last resource that I'll share is https://fediverse.info. This is like the tr
|
|||
|
||||
## Conclusion
|
||||
|
||||
The lack of recommendation engine has driven the Mastodon community to create new and novel ways of discovering new people. It wouldn't surprise me if within the year a whole new set of tools are made available to us. Though I have to say, the default of not showing unfollowed accounts in our feeds is really nice. It means that our Mastodon experience is truly unique to us, and that we can curate them to our liking.
|
||||
The lack of recommendation engine has driven the Mastodon community to create new and novel ways of discovering new people. It wouldn't surprise me if within the year a whole new set of tools are made available to us. Though I have to say, the default of not showing unfollowed accounts in our feeds is really nice. It means that our Mastodon experience is truly unique to us, and that we can curate them to our liking.
|
||||
|
|
@ -1,10 +1,11 @@
|
|||
---
|
||||
title: "Finding Cuda Errors"
|
||||
date: 2022-03-08T10:53:43-05:00
|
||||
date: 2022-03-08 10:53:43-05:00
|
||||
draft: false
|
||||
tags: []
|
||||
math: false
|
||||
medium_enabled: true
|
||||
medium_post_id: a07f460ce553
|
||||
tags: []
|
||||
title: Finding Cuda Errors
|
||||
---
|
||||
|
||||
When cuda fails, it fails silently. To combat this, I have gotten into a habit of checking for a failed status for every cuda memory allocation, kernel execution etc. The following is a C++ macro I wrote that checks if a previous cuda call has failed and then prints the current line and file of the macro.
|
||||
|
|
|
|||
|
|
@ -1,10 +1,12 @@
|
|||
---
|
||||
title: "Which commit broke the build? Using Git Bisect"
|
||||
date: 2022-05-03T01:15:55-04:00
|
||||
date: 2022-05-03 01:15:55-04:00
|
||||
draft: false
|
||||
tags: ["Git"]
|
||||
math: false
|
||||
medium_enabled: true
|
||||
medium_post_id: 1ef72415ab75
|
||||
tags:
|
||||
- Git
|
||||
title: Which commit broke the build? Using Git Bisect
|
||||
---
|
||||
|
||||
Lets imagine a scenario where in the latest merge a test
|
||||
|
|
|
|||
|
|
@ -1,15 +1,16 @@
|
|||
---
|
||||
title: "Git Diff by Words"
|
||||
date: 2022-02-04T21:23:30-05:00
|
||||
date: 2022-02-04 21:23:30-05:00
|
||||
draft: false
|
||||
tags: ["Git"]
|
||||
math: false
|
||||
medium_enabled: true
|
||||
medium_post_id: 1fa7eb8ae0f5
|
||||
tags:
|
||||
- Git
|
||||
title: Git Diff by Words
|
||||
---
|
||||
|
||||
When working in LaTex it is often more useful to see the changes by words as opposed to lines. Luckily, it's pretty easy to view this in the terminal.
|
||||
|
||||
```bash
|
||||
git diff --color-words
|
||||
```
|
||||
|
||||
```
|
||||
|
|
@ -1,10 +1,12 @@
|
|||
---
|
||||
title: "Git Partial Clones"
|
||||
date: 2022-02-07T17:07:08-05:00
|
||||
date: 2022-02-07 17:07:08-05:00
|
||||
draft: false
|
||||
tags: ["Git"]
|
||||
math: false
|
||||
medium_enabled: true
|
||||
medium_post_id: 95d4bd60af72
|
||||
tags:
|
||||
- Git
|
||||
title: Git Partial Clones
|
||||
---
|
||||
|
||||
I recently was introduced to [Sparse Directories in SVN](https://svnbook.red-bean.com/en/1.8/svn.advanced.sparsedirs.html). In SVN, you can initially clone a repository and have it be empty until you checkout the specific files needed. I wondered if I can do the same with `git`. For the *tl;dr* skip to the conclusion section.
|
||||
|
|
@ -119,5 +121,4 @@ You can even set it so that a specific folder is shown at the root of your direc
|
|||
|
||||
```bash
|
||||
git sparse-checkout set FOLDERNAME
|
||||
```
|
||||
|
||||
```
|
||||
|
|
@ -1,10 +1,11 @@
|
|||
---
|
||||
title: "Intensional Logic Extends First Order"
|
||||
date: 2022-02-26T20:33:38-05:00
|
||||
date: 2022-02-26 20:33:38-05:00
|
||||
draft: false
|
||||
tags: []
|
||||
math: true
|
||||
medium_enabled: true
|
||||
medium_post_id: 8e07b68aa8f5
|
||||
tags: []
|
||||
title: Intensional Logic Extends First Order
|
||||
---
|
||||
|
||||
The second brightest object in the sky is known as the morgensteorra (morning star) and æfensteorra (evening star). Later on this object became known as Venus. [(Wikipedia)](https://en.wikipedia.org/wiki/Venus_in_culture)
|
||||
|
|
@ -21,4 +22,4 @@ In first order logic, we can then deduce the following:
|
|||
$$
|
||||
B(\text{morgensteorra} = \text{venus})
|
||||
$$
|
||||
But does that make sense? It is possible to hold a belief that Venus is the evening star while not holding a belief that Venus is the morning star. Therefore, we cannot treat belief as a traditional relation symbol. Issues like these give birth to intensional reasoning and from that modal logic.
|
||||
But does that make sense? It is possible to hold a belief that Venus is the evening star while not holding a belief that Venus is the morning star. Therefore, we cannot treat belief as a traditional relation symbol. Issues like these give birth to intensional reasoning and from that modal logic.
|
||||
|
|
@ -1,10 +1,12 @@
|
|||
---
|
||||
title: "JUnit Script"
|
||||
date: 2022-02-26T20:02:53-05:00
|
||||
date: 2022-02-26 20:02:53-05:00
|
||||
draft: false
|
||||
tags: ["Java"]
|
||||
math: false
|
||||
medium_enabled: true
|
||||
medium_post_id: 7ca070e54955
|
||||
tags:
|
||||
- Java
|
||||
title: JUnit Script
|
||||
---
|
||||
|
||||
Running a JUnit test on the terminal is a little annoying. Here is a quick script to make it easier, make it executable and drop it in [`~/.local/bin`](/blog/customexec/) for easy use.
|
||||
|
|
@ -32,4 +34,4 @@ java -cp "$JUNIT_PATH":. org.junit.runner.JUnitCore "$1"
|
|||
|
||||
```
|
||||
|
||||
Replacing `JUNIT_PATH` with the location of the `junit4.jar` on your system.
|
||||
Replacing `JUNIT_PATH` with the location of the `junit4.jar` on your system.
|
||||
|
|
@ -1,10 +1,12 @@
|
|||
---
|
||||
title: "LaTex List Labels"
|
||||
date: 2022-01-16T23:17:51-05:00
|
||||
date: 2022-01-16 23:17:51-05:00
|
||||
draft: false
|
||||
tags: ["LaTex"]
|
||||
math: false
|
||||
medium_enabled: true
|
||||
medium_post_id: b978cbe10da3
|
||||
tags:
|
||||
- LaTex
|
||||
title: LaTex List Labels
|
||||
---
|
||||
|
||||
A quick tip that I recently learned is that the symbols in a LaTex list item is changeable. In fact, the following technique works for both the `enumerate` and `itemize` environments.
|
||||
|
|
@ -22,4 +24,4 @@ A quick tip that I recently learned is that the symbols in a LaTex list item is
|
|||
\end{enumerate}
|
||||
```
|
||||
|
||||
|
||||
|
||||
|
|
@ -1,10 +1,12 @@
|
|||
---
|
||||
title: "Multi-Column slides in Beamer"
|
||||
date: 2022-01-26T23:29:31-05:00
|
||||
date: 2022-01-26 23:29:31-05:00
|
||||
draft: false
|
||||
tags: ["LaTex"]
|
||||
math: false
|
||||
medium_enabled: true
|
||||
medium_post_id: cd19b12d502f
|
||||
tags:
|
||||
- LaTex
|
||||
title: Multi-Column slides in Beamer
|
||||
---
|
||||
|
||||
When creating slides in LaTex Beamer, it can be frustrating to not have the easy ability to drag text boxes around. Luckily, creating a multi-column layout in Beamer is not difficult! Below is an example of a two column slide layout, where the left side is a bulleted list, and the right side is an image:
|
||||
|
|
@ -40,4 +42,4 @@ When creating slides in LaTex Beamer, it can be frustrating to not have the easy
|
|||
\end{document}
|
||||
```
|
||||
|
||||
|
||||
|
||||
|
|
@ -1,10 +1,12 @@
|
|||
---
|
||||
title: "Networkx Random Sample Graph"
|
||||
date: 2022-04-07T19:48:12-04:00
|
||||
date: 2022-04-07 19:48:12-04:00
|
||||
draft: false
|
||||
tags: ["Python"]
|
||||
math: false
|
||||
medium_enabled: true
|
||||
medium_post_id: e5c5330cc9a7
|
||||
tags:
|
||||
- Python
|
||||
title: Networkx Random Sample Graph
|
||||
---
|
||||
|
||||
I've been working on several algorithms in `networkx`. In order to speed up testing, especially on large graphs, I've been randomly sampling portions of the original graph. The best way I've found to do this is through the following python snippet:
|
||||
|
|
@ -24,4 +26,4 @@ random_nodes = random.sample(list(G.nodes), SAMPLE_SIZE)
|
|||
G_sample = G.subgraph(random_nodes)
|
||||
```
|
||||
|
||||
However, only considering the nodes when sampling makes it highly likely that the subgraph will significantly less edges. This results in a mostly disconnected subgraph and a loss of information. Sampling the edges prevents this issue at the expense of not capturing single nodes not connected to anything else.
|
||||
However, only considering the nodes when sampling makes it highly likely that the subgraph will significantly less edges. This results in a mostly disconnected subgraph and a loss of information. Sampling the edges prevents this issue at the expense of not capturing single nodes not connected to anything else.
|
||||
|
|
@ -1,10 +1,12 @@
|
|||
---
|
||||
title: "Tree-based Plots in NetworkX"
|
||||
date: 2022-01-02T20:36:41-05:00
|
||||
date: 2022-01-02 20:36:41-05:00
|
||||
draft: false
|
||||
tags: ["Python"]
|
||||
math: false
|
||||
medium_enabled: true
|
||||
medium_post_id: c8cffb5c4078
|
||||
tags:
|
||||
- Python
|
||||
title: Tree-based Plots in NetworkX
|
||||
---
|
||||
|
||||
A graph in D3 and NetworkX can be represented as a JSON file.
|
||||
|
|
@ -51,4 +53,4 @@ plt.show()
|
|||
|
||||
Given the JSON from the top of the post it'll produce:
|
||||
|
||||
|
||||
|
||||
|
|
@ -1,10 +1,11 @@
|
|||
---
|
||||
title: "Reducing Network Bandwidth in Nginx with Gzip"
|
||||
date: 2022-02-04T20:06:50-05:00
|
||||
date: 2022-02-04 20:06:50-05:00
|
||||
draft: false
|
||||
tags: []
|
||||
math: false
|
||||
medium_enabled: true
|
||||
medium_post_id: ebd510dbf691
|
||||
tags: []
|
||||
title: Reducing Network Bandwidth in Nginx with Gzip
|
||||
---
|
||||
|
||||
Browsers that support gzip compression send the header `Accept-Encoding: gzip` in its request and if the webserver supports gzip, then it can send the website data back compressed. Though recently I learned that this isn't setup by default in `nginx`! This post will go over the configuration you'll need to add to `/etc/nginx/nginx.conf` in order to support `gzip` compression. From my experience, enabling `gzip` compression reduced network traffic by 1/4.
|
||||
|
|
@ -54,4 +55,4 @@ What each of the commands do:
|
|||
| `gzip_buffers` | Sets the `*number*` and `*size*` of buffers used to compress a response. By default, the buffer size is equal to one memory page. This is either 4K or 8K, depending on a platform. |
|
||||
| `gzip_http_version` | Sets the minimum HTTP version of a request required to compress a response. |
|
||||
| `gzip_min_length` | Sets the minimum length of a response that will be gzipped. The length is determined only from the “Content-Length” response header field. |
|
||||
| `gzip_types` | Enables gzipping of responses for the specified MIME types in addition to “`text/html`”. The special value “`*`” matches any MIME type (0.8.29). Responses with the “`text/html`” type are always compressed. |
|
||||
| `gzip_types` | Enables gzipping of responses for the specified MIME types in addition to “`text/html`”. The special value “`*`” matches any MIME type (0.8.29). Responses with the “`text/html`” type are always compressed. |
|
||||
|
|
@ -1,10 +1,12 @@
|
|||
---
|
||||
title: "Non-Root Systemd Scripts"
|
||||
date: 2022-03-15T23:55:19-04:00
|
||||
date: 2022-03-15 23:55:19-04:00
|
||||
draft: false
|
||||
tags: ["Linux"]
|
||||
math: false
|
||||
medium_enabled: true
|
||||
medium_post_id: f527d0f12fb8
|
||||
tags:
|
||||
- Linux
|
||||
title: Non-Root Systemd Scripts
|
||||
---
|
||||
|
||||
I know of two ways to run systemd services not as root. They both have their pros and cons associated with them.
|
||||
|
|
@ -105,4 +107,4 @@ From [FreeDesktop](https://www.freedesktop.org/software/systemd/man/loginctl.htm
|
|||
|
||||
For a multi-user system, one can argue that linger is not a good property to have because users can then have a lot of processes spawned and persistent even if they are not using the system. In that case, having an admin individually add systemd unit files using the first approach makes sense.
|
||||
|
||||
The second approach has the benefit that the admin only has to set the linger property of the user once, and then the user can create as many systemd unit services as they'd like.
|
||||
The second approach has the benefit that the admin only has to set the linger property of the user once, and then the user can create as many systemd unit services as they'd like.
|
||||
|
|
@ -1,10 +1,12 @@
|
|||
---
|
||||
title: "OpenMPI on Fedora"
|
||||
date: 2022-05-03T00:52:38-04:00
|
||||
date: 2022-05-03 00:52:38-04:00
|
||||
draft: false
|
||||
tags: ["Fedora"]
|
||||
math: false
|
||||
medium_enabled: true
|
||||
medium_post_id: 1c0c98b99e8a
|
||||
tags:
|
||||
- Fedora
|
||||
title: OpenMPI on Fedora
|
||||
---
|
||||
|
||||
To use use the OpenMPI compilers (mpicc, mpic++, etc.) and mpirun
|
||||
|
|
@ -33,4 +35,4 @@ module load mpi/openmpi-x86_64
|
|||
Finally, with these changes you can use the compiler
|
||||
tools and runner. Do note that you'll have to source
|
||||
and load the OpenMPI module for every shell you open
|
||||
unless you add it within `$HOME/.bashrc`.
|
||||
unless you add it within `$HOME/.bashrc`.
|
||||
|
|
@ -1,10 +1,11 @@
|
|||
---
|
||||
title: "Permission Denied: Writing to Privileged Locations"
|
||||
date: 2022-04-07T20:09:40-04:00
|
||||
date: 2022-04-07 20:09:40-04:00
|
||||
draft: false
|
||||
tags: []
|
||||
math: false
|
||||
medium_enabled: true
|
||||
medium_post_id: 7d133254be3f
|
||||
tags: []
|
||||
title: 'Permission Denied: Writing to Privileged Locations'
|
||||
---
|
||||
|
||||
Perhaps you've tried something like the following:
|
||||
|
|
@ -27,4 +28,4 @@ Instead consider the following:
|
|||
echo "hi" | sudo tee /etc/test
|
||||
```
|
||||
|
||||
The command `tee` takes whatever is in standard in, and writes it to the filename specified. Since we applied `sudo ` to the `tee` command, it now has the necessary permissions to write to a privileged location.
|
||||
The command `tee` takes whatever is in standard in, and writes it to the filename specified. Since we applied `sudo ` to the `tee` command, it now has the necessary permissions to write to a privileged location.
|
||||
|
|
@ -1,10 +1,12 @@
|
|||
---
|
||||
title: "Program Verification with Hoare Logic and Dafny"
|
||||
date: 2022-02-05T00:06:42-05:00
|
||||
date: 2022-02-05 00:06:42-05:00
|
||||
draft: false
|
||||
tags: ["Formal Methods"]
|
||||
math: false
|
||||
medium_enabled: true
|
||||
medium_post_id: 4cf20da7d8a4
|
||||
tags:
|
||||
- Formal Methods
|
||||
title: Program Verification with Hoare Logic and Dafny
|
||||
---
|
||||
|
||||
In the recitations that I'm giving for [Principles of Software](/ta/spring2022/csci2600/), we are going over reasoning through code using Hoare Logic and the program verifier Dafny. Microsoft Research designed Dafny to be similar to writing imperative code. The main difference is that you need to supply (pre/post)-conditions and the code to verify. Here's an example of how to reason about a simple statement by hand:
|
||||
|
|
@ -55,9 +57,4 @@ method abs(x: int) returns (y: int)
|
|||
assert y == x;
|
||||
}
|
||||
}
|
||||
```
|
||||
|
||||
|
||||
|
||||
|
||||
|
||||
```
|
||||
|
|
@ -1,10 +1,12 @@
|
|||
---
|
||||
title: "Loop Invariants"
|
||||
date: 2022-02-26T19:17:21-05:00
|
||||
date: 2022-02-26 19:17:21-05:00
|
||||
draft: false
|
||||
tags: ["Formal Methods"]
|
||||
math: true
|
||||
medium_enabled: true
|
||||
medium_post_id: 395dbef73491
|
||||
tags:
|
||||
- Formal Methods
|
||||
title: Loop Invariants
|
||||
---
|
||||
|
||||
Loop invariants are one of the more complicated concepts in Hoare Logic. One might say that we can simply unroll the loop, but there are many loops where we don't know the number of iterations in advanced. Therefore, we often consider what is called the loop invariant. These are statements that are true before, during, and after the loop. A useful loop invariant in combination of the negation of the loop condition can also prove some postcondition. Lets look at a multiplication example
|
||||
|
|
@ -89,4 +91,4 @@ $$
|
|||
(x_n == a * p_n) \wedge (p \ge b) \wedge (p \le b) \\\\
|
||||
\implies (x_n == a * p_n) \wedge (p == b) \\\\
|
||||
\implies (x_n == a * b) \checkmark \\
|
||||
$$
|
||||
$$
|
||||
|
|
@ -1,10 +1,12 @@
|
|||
---
|
||||
title: "Rclone: The Swiss Army Knife of Cloud Storage"
|
||||
date: 2022-04-10T22:05:46-04:00
|
||||
date: 2022-04-10 22:05:46-04:00
|
||||
draft: false
|
||||
tags: ["Backup"]
|
||||
math: false
|
||||
medium_enabled: true
|
||||
medium_post_id: ee4c7c4d2966
|
||||
tags:
|
||||
- Backup
|
||||
title: 'Rclone: The Swiss Army Knife of Cloud Storage'
|
||||
---
|
||||
|
||||
[Rclone](https://rclone.org/) is a terminal application that manages remote storage. It supports the standard remote filesystem protocols such as S3, (S)FTP, WebDAV as well as easy to use integration with cloud providers such as Amazon S2, Backblaze B2, Google Drive, and others. We can even perform regular backups using [Restic](https://restic.net/) with Rclone underneath. With all these features, it's easy to configure all remote storage within Rclone and then synchronize the configuration file across machines.
|
||||
|
|
@ -221,4 +223,4 @@ WantedBy=default.target
|
|||
|
||||
## Conclusion
|
||||
|
||||
This is just the tip of the iceburg when it comes to Rclone! Another popular feature is automatically [encrypting](https://rclone.org/crypt/) files uploaded to cloud storage. Once you have everything configured to your liking, you can synchronize the file under `~/.config/rclone/rclone.conf` across all your machines.
|
||||
This is just the tip of the iceburg when it comes to Rclone! Another popular feature is automatically [encrypting](https://rclone.org/crypt/) files uploaded to cloud storage. Once you have everything configured to your liking, you can synchronize the file under `~/.config/rclone/rclone.conf` across all your machines.
|
||||
|
|
@ -1,10 +1,13 @@
|
|||
---
|
||||
title: "Rebuild Kernel Modules with Akmods"
|
||||
date: 2022-02-04T19:37:04-05:00
|
||||
date: 2022-02-04 19:37:04-05:00
|
||||
draft: false
|
||||
tags: ["Linux", "Fedora"]
|
||||
math: false
|
||||
medium_enabled: true
|
||||
medium_post_id: ab0e2a2ce3b3
|
||||
tags:
|
||||
- Linux
|
||||
- Fedora
|
||||
title: Rebuild Kernel Modules with Akmods
|
||||
---
|
||||
|
||||
Akmods is the Fedora/Red Hat way of managing kernel modules. In Ubuntu, this is `dkms`. If you're like me and force reboot shortly after performing an update, then you might have not given akmods enough time to compile any extra kernel modules (for example: Nvidia). This meant that I had to boot into an older kernel to try to fix the problem....
|
||||
|
|
@ -26,4 +29,4 @@ to trigger the rebuild.
|
|||
Though the better solution is to avoid this problem to begin with.
|
||||
If you `reboot` not as root, then systemd will check to see if
|
||||
any process is inhibiting the poweroff. If that's the case,
|
||||
wait patiently and don't type `sudo reboot`.
|
||||
wait patiently and don't type `sudo reboot`.
|
||||
|
|
@ -1,10 +1,12 @@
|
|||
---
|
||||
title: "Rootless Docker-Compose with Podman"
|
||||
date: 2022-01-29T20:21:11-05:00
|
||||
date: 2022-01-29 20:21:11-05:00
|
||||
draft: false
|
||||
tags: ["Containers"]
|
||||
math: false
|
||||
medium_enabled: true
|
||||
medium_post_id: d31b7e2a688c
|
||||
tags:
|
||||
- Containers
|
||||
title: Rootless Docker-Compose with Podman
|
||||
---
|
||||
|
||||
One of the benefits of Podman over Docker is that it can run daemon-less and without root. However, `docker-compose` is by far my favorite way to create and maintain containers. Luckily, the Podman folks emulated the Docker CLI so that `docker-compose` works well with Podman!
|
||||
|
|
@ -71,4 +73,4 @@ Now we can run `docker-compose`!
|
|||
docker-compose ps
|
||||
```
|
||||
|
||||
[^1]: https://bugzilla.redhat.com/show_bug.cgi?id=2125878
|
||||
[^1]: https://bugzilla.redhat.com/show_bug.cgi?id=2125878
|
||||
|
|
@ -1,10 +1,12 @@
|
|||
---
|
||||
title: "Mount Object Storage Locally using S3 Fuse"
|
||||
date: 2022-02-04T23:39:25-05:00
|
||||
date: 2022-02-04 23:39:25-05:00
|
||||
draft: false
|
||||
tags: ["Storage"]
|
||||
math: false
|
||||
medium_enabled: true
|
||||
medium_post_id: a6add7081378
|
||||
tags:
|
||||
- Storage
|
||||
title: Mount Object Storage Locally using S3 Fuse
|
||||
---
|
||||
|
||||
On most cloud providers, object storage is cheaper than paying for the equivalent size in block storage. Using FUSE, we can mount S3 compatible object storage with the command `s3fs`. Do note, that there are a few downsides with mounting object storage as documented on their [README](https://github.com/s3fs-fuse/s3fs-fuse/blob/master/README.md):
|
||||
|
|
@ -61,5 +63,4 @@ To mount automatically during reboot, add the following to `/etc/fstab`:
|
|||
bucketname /mnt/mountpoint fuse.s3fs _netdev,allow_other,url=https://us-east-1.linodeobjects.com 0 0
|
||||
```
|
||||
|
||||
After editing `/etc/fstab` you can run `sudo mount -a` in order for it to load and mount any new entries.
|
||||
|
||||
After editing `/etc/fstab` you can run `sudo mount -a` in order for it to load and mount any new entries.
|
||||
|
|
@ -1,10 +1,12 @@
|
|||
---
|
||||
title: "Aliases for Snaps and Flatpaks"
|
||||
date: 2022-02-04T20:29:04-05:00
|
||||
date: 2022-02-04 20:29:04-05:00
|
||||
draft: false
|
||||
tags: ["Linux"]
|
||||
math: false
|
||||
medium_enabled: true
|
||||
medium_post_id: 5f51916cafd3
|
||||
tags:
|
||||
- Linux
|
||||
title: Aliases for Snaps and Flatpaks
|
||||
---
|
||||
|
||||
Sometimes snap packages and flatpaks come with unmemorable names to run their commands. Luckily for snap packages, you can easily alias them. For example, to rename the .NET executable to `dotnet` run.
|
||||
|
|
@ -34,5 +36,4 @@ done
|
|||
|
||||
# main
|
||||
launch_app "${1}"
|
||||
```
|
||||
|
||||
```
|
||||
|
|
@ -1,10 +1,13 @@
|
|||
---
|
||||
title: "Configuring a SOCKS5 Proxy"
|
||||
date: 2022-01-03T10:41:09-05:00
|
||||
date: 2022-01-03 10:41:09-05:00
|
||||
draft: false
|
||||
tags: ["Networking", "SSH"]
|
||||
math: false
|
||||
medium_enabled: true
|
||||
medium_post_id: ed506a09a2b2
|
||||
tags:
|
||||
- Networking
|
||||
- SSH
|
||||
title: Configuring a SOCKS5 Proxy
|
||||
---
|
||||
|
||||
A SOCKS5 proxy allows you to have network traffic as if it was coming from the proxy server as opposed to your local client. You can easily set it up using SSH from your local machine.
|
||||
|
|
@ -31,5 +34,4 @@ To do this you will need proxychains installed. On Ubuntu, the package is called
|
|||
|
||||
```
|
||||
socks5 127.0.0.1 1337
|
||||
```
|
||||
|
||||
```
|
||||
|
|
@ -1,10 +1,11 @@
|
|||
---
|
||||
title: "Terminal Output To SVG Animations"
|
||||
date: 2022-01-17T10:14:22-05:00
|
||||
date: 2022-01-17 10:14:22-05:00
|
||||
draft: false
|
||||
tags: []
|
||||
math: false
|
||||
medium_enabled: true
|
||||
medium_post_id: ad8c533610b5
|
||||
tags: []
|
||||
title: Terminal Output To SVG Animations
|
||||
---
|
||||
|
||||
With [`termtosvg`](https://github.com/nbedos/termtosvg) made by Nicolas Bedo we can make SVG animations from terminal output in the style of [asciinema](https://asciinema.org/). To install use [pipx](/blog/managepythonapps/).
|
||||
|
|
@ -53,4 +54,4 @@ termtosvg -s
|
|||
|
||||
Here is an example of an animation I made with this tool:
|
||||
|
||||
|
||||
|
||||
|
|
@ -1,10 +1,12 @@
|
|||
---
|
||||
title: "Why I PESOS From Mastodon"
|
||||
date: 2022-05-15T18:36:16-04:00
|
||||
date: 2022-05-15 18:36:16-04:00
|
||||
draft: false
|
||||
tags: ["Mastodon"]
|
||||
math: false
|
||||
medium_enabled: true
|
||||
medium_post_id: 56667e2c305
|
||||
tags:
|
||||
- Mastodon
|
||||
title: Why I PESOS From Mastodon
|
||||
---
|
||||
|
||||
The [IndieWeb movement](https://indieweb.org) is focused on people controlling their own experience on the web. They accomplish this by encouraging everyone to have their own personal website. Followers can then either subscribe to the [RSS feed](https://aboutfeeds.com/) and provide [Webmentions](https://indieweb.org/Webmention) back to the post, or they can interact with the material directly on another website like [Mastodon](https://joinmastodon.org/).
|
||||
|
|
@ -26,4 +28,4 @@ The main advantage of PESOS is that my primary interaction with Mastodon is thro
|
|||
|
||||
## Disadvantages to PESOS
|
||||
|
||||
The primary way that people are following me is within Mastodon itself as opposed to the RSS feed on my website. This means that if the Mastodon server ever abruptly goes down, then I will lose contact with all of my followers. Mastodon, however, does have some protection against this as a federated network. Mainly the ability to [move your profile](https://docs.joinmastodon.org/user/moving/) from one Mastodon instance to another. This gives me a more secure feeling that Mastodon itself won't shut down in the near future.
|
||||
The primary way that people are following me is within Mastodon itself as opposed to the RSS feed on my website. This means that if the Mastodon server ever abruptly goes down, then I will lose contact with all of my followers. Mastodon, however, does have some protection against this as a federated network. Mainly the ability to [move your profile](https://docs.joinmastodon.org/user/moving/) from one Mastodon instance to another. This gives me a more secure feeling that Mastodon itself won't shut down in the near future.
|
||||
Loading…
Reference in a new issue