How the shit do I remember things like "Pumas on Hoverbikes is at monkeybagel.com" but I have to remind myself to eat lunch as a separate step from making lunch?
For some reason, I decided that my video game tonight would be to screw around with ansible in github actions deploying to a brand new ubuntu cloud host.
@wohali iā¦ have ended up doing that before š
I reworked my blog so it's easier for me to post to it, so it publishes more often.
Turns out I never optimized the publishing process to be very incremental, so it's reuploading the entire site to S3 every time with `aws s3 sync`
Turns out that doubled my S3 bill already. Granted, that's from $3 to $6, but I don't like a 2x multiplier on anything š
Looks like `aws s3 sync` isn't at all an analogue for `rsync` whatsoever š
Trying to figure out who's singing with Peter Murphy at about 1:39 in "Silver Shade" - I *think* it's Trent Reznor? He almost sounds like he's doing KISS karaoke
@Mossop Yeah, "$0.005 per 1,000 PUT, COPY, POST, or LIST requests"
That adds up when you're being dumb š
The main excuse is that I think I wrote this github action workflow like 7 years ago, back when I managed to blog maybe 6 times a year. So it wasn't a problem until I wanted to, like, you know, use it
@cobweb This is a surprisingly star-studded album!
Dang, the personnel listing on this new Peter Murphy album reads like a Yes reunion
@Mossop Yeah, this is mainly inertia and cobwebs I'm blowing off. I've had a site hosted on S3 since probably... oof... 2007
@socketwench An eye patch helps too
@Mossop Hmm, I do already use Backblaze for offsite backups š¤
@anticdent I'm on probably a 4th or 5th listen through and I've been pretty pleased with it. I think it's one of the best things he's done in a long while
@randomgeek "join our discord"
@undead Yeah, I've taken to like a virtual desktop per task, and sometimes that's a desktop across 3 monitors or so.
Nice to flip entirely between context switches. Also have lots and lots of browser windows each devoted to one research wandering
I am continually frustrated by macOS' handling of this, since they really seem to think you actually just want to narrow down to one app at a time ever
Woot! I love voting here, so civilized #pdx
Also really digging the new VNV Nation disc "Construct" - but then I always dig VNV Nation.
I've had surly rivetheads try to convince me they're a Christian band. But, you know, you can have some electronic music that's sappy and hopeful. Also cringingly melodramatic and occasionally uplifting. It's not a sin
I don't go to many shows anymore, but I did see VNV Nation live. Was right up front. Not ashamed to admit I'm an enormous sucker for Perpetual. He did like a 15 minute long extension of the chorus with everyone. I will admit my eyes were not dry
And I'm always made verklempt if I really think about the lyrics of "The Farthest Star" - I mean this really gets me in the Star Trek believing kidsoul...
We possess the power
If this should start to fall apart
To mend divides, to change the world
To reach the farthest star
Anyway, I don't post on Livejournal anymore, so these fits of lyric posting come out occasionally
@rckenned Yeah, all the files get regenerated in a fresh filesystem in a stateless github action. Ideally it should be using a hash set of the files, which some other tools do.
I've tried the size-only option before and it left me with a bunch of files that were different yet coincidentally the same size. That was especially annoying for this client-side search thing I'm using (Pagefind) that breaks the indices up into equal sized chunks
Oh and I forget that they let me know when it's been *counted* too
Playing with switching to rclone for my s3 blog publishing. Wrote a quick github workflow, let's see if it works
(it automatically deletes all blog files from S3)
Nope, that didn't work.
Good thing I generate the whole site from scratch every time. š (oh and that I downloaded a backup first)
The good news is that I think I fixed the rclone workflow. Only took 6 minutes to re-upload everything and then only 1 minute for a subsequent publish. So, I think that's good news?
Heh, I've got a few oooooold Amazon S3 buckets
That s3wiki bucket was from a far more innocent time when I thought it was clever to hook up a public-writable S3 bucket to a javascript wiki frontend that could edit itself
https://blog.lmorchard.com/2006/04/21/an-s3-ajax-wiki/index.html
Also while doing some personal infra late-spring cleaning, I realized I had a bunch of my S3 buckets hosting static sites writing access logs to another bucket.
I guess I'd intended to actually do something with those someday? I never did. At least I had presence of mind to stick a 90-day deletion lifecycle on it. Still, there are so many objects in that bucket š±
Turns out S3's idea of access logging is writing a 400b timestamped file every few seconds, forever? Like log rotation but measured in RPMs
Chat, am I hosed?
