From b44ac1afd0a50d0db9dab8bbdb4b9bd0b9b54d33 Mon Sep 17 00:00:00 2001
From: Kyle Isom <kyle@imap.cc>
Date: Mon, 28 Feb 2022 21:45:54 -0800
Subject: [PATCH] Stub out drafts for the next two sets of stream notes.

---
 content/posts/stream-0x03.md    | 30 ++++++++++++++++++++++++++++++
 content/posts/stream-0x04.md.md | 13 +++++++++++++
 2 files changed, 43 insertions(+)
 create mode 100644 content/posts/stream-0x03.md
 create mode 100644 content/posts/stream-0x04.md.md

diff --git a/content/posts/stream-0x03.md b/content/posts/stream-0x03.md
new file mode 100644
index 0000000..bb0f24f
--- /dev/null
+++ b/content/posts/stream-0x03.md
@@ -0,0 +1,30 @@
+Title: Stream 0x03
+Slug: stream-0x03
+Date: 2022-03-01
+Modified: 2022-03-01
+Category: 
+Tags: stream
+Authors: kyle
+Summary: Stream notes for tonight's stream.
+Status: draft
+
+Tonight's work focused on adding in a mirror between local storage and
+a remote S3 (Minio) instance. The basic flow goes something like:
+
+1. Fetch the object described
+2. Store the object locally
+3. Throw the object ID onto the work queue
+4. The work queue pulls an item off the queue.
+5. Load the object from local storage.
+6. If there's no error, go look for the next item.
+7. If the job doesn't have a backoff set, initialize a new backoff.
+8. Launch a goroutine that waits as long as the backoff says before
+   putting the item back on the work queue.
+
+Next stream (2021/03/03), we'll look at designing the artifact, maybe
+working on some of the protobuf definitions.
+
+### References
+
+* [cloudflare/backoff](https://github.com/cloudflare/backoff/) - software
+  I wrote at a previous job.
diff --git a/content/posts/stream-0x04.md.md b/content/posts/stream-0x04.md.md
new file mode 100644
index 0000000..9c7906e
--- /dev/null
+++ b/content/posts/stream-0x04.md.md
@@ -0,0 +1,13 @@
+Title: stream-0x04.md
+Slug: stream-0x04.md
+Date: 2022-03-03
+Modified: 2022-03-03
+Category: 
+Tags: stream
+Authors: kyle
+Summary: 
+Status: draft
+
+In tonight's stream, we looked at building out the artifact data
+type. This started with reviewing the artifact spec, then translating
+that into protobufs. This also required the SQL schema to be developed.