iiinova — where your digital self keeps evolving

... but that could've been recoverable, even if it took some time and investment. And even then, could minimize it by sorta cutting the losses of Destiny 2 and starting fresh with Destiny 3 as a kinda "middle ground".

210

Like... I realize that this isn't JUST about Sony making a bad long-term decision for short-term gains... former Bungie leadership really fucked up here, and over-sold things to Sony and meddled in the development that led to player frustrations...

130

The right move would be to start tapering off and winding down Destiny 2 while ramping up for a clean break to Destiny 3. Tie off the story of Destiny 2 with a lead-in to Destiny 3. It's okay to have a gap... but just have a PLAN.

I won't deny that Destiny 2 was probably "more expensive to maintain" than it was really worth, that's probably true. But just canceling the franchise with no plans is a poor long-term strategy, and only balances the quarterly budget on paper.

And just looking at these trends, Marathon's gonna keep declining (probably until the free week is over, then go back to it's normal levels)... and Destiny 2 will keep increasing and probably spike higher than Marathon's launch numbers when the final update drops. Again... yikes...

I'm not gonna SAY that Sony/PlayStation made a mistake... but... The game that recently announced end-of-service with one week until the final update is currently, on it's WORST platform (Steam/PC), outperforming Marathon's "free week" on it's BEST platform (Steam/PC)... ... yikes.

I guess that means the space utilized by the database is ~7.4x by migrating from SQLite to MariaDB/Galera... Ideal would've been 3x... so 2.46x above "ideal"

000

Actually, might be better to compare from the SQL dump file size, rather than the SQLite DB... 34GB. So... 28GB SQLite -> 34GB SQL dump -> 69GB MariaDB -> 207GB Galera replication -> 828GB 4x Ceph replication Technically, the 28GB SQLite file would also be 4x replicated, though (112GB)... meh.

100

But mostly... while 28GB->69GB moving from SQLite to MariaDB is shocking in it's own right (yay indexes!)... just counting up the multiple layers of replication and seeing it balloon to nearly 1TB is just... 🤯 (even though I understand it, SEEING it is something else...)

100

And TECHNICALLY it's resilient and fault-tolerant... but since Home Assistant itself is still the single-point-of-failure, that's pretty pointless. In theory, though... easier to snapshot/backup the DB now, though.... so... there's that.

100

I mean, it performs better/faster for reads now than it did before (technically, reads can be spread across the 3 hosts, and reads can leverage parallel reads from 4 OSDs at once... per host... in theory)... But oooof... that's a lot of overhead to get there.

100

Bryan Wyatt @brwyatt.net · 28/05/2026

Always fun when you do the math on something simple... I migrated a 28GB SQLite Home Assistant DB to a 3-host MariaDB/Galera cluster backed by a 4x replicated Ceph RBD pool. At 69GB in MariaDB (yay indexes?), x3 hosts, x4 replicated... 828GB 😰

100

That, and a general lack of energy for a couple weeks (before things adjust), is the worst part of starting a calorie reduction plan. It suuuuuucks. Gotta stay strong, though! You can do it!

000

Bryan Wyatt @brwyatt.net · 25/05/2026

Probably meaningless... but... well... it's the only thing I can think to do right now: c.org/kfZM94Wvwx

c.org

Sign the Petition

Petition Sony to develop Destiny 3

000

Bryan Wyatt @brwyatt.net · 25/05/2026

I know it's kinda silly to feel this kind of "loss" for a game... But really it was more than that. And it was always "there", and I could catch up on all I missed, and exist in this world for a time. My dad and I used to play together a lot (on Stadia!)... Just kinda... Sucks.

000

Bryan Wyatt @brwyatt.net · 25/05/2026

I don't know what to think right now... Destiny 2 was a game - a world - that I always kept coming back to. There was something about it that is hard to articulate, even through the missteps. For right now, I'm just allowing myself to be sad and a little angry.

100

Bryan Wyatt @brwyatt.net · 21/04/2026

So... I decided to check in on my UPSs for the first time in... oh... 2+ years. 1) There's a firmware update (yay!) 2) I have an TLS cert manager container that can push TLS certs to devices now! (HTTPS! YAY!) 3) I have an LDAP server now, and these can use it! Exciting weekend ahead, I think!

000

Reposted by Bryan Wyatt

Savestate Comic @savestatecomic.com · 08/04/2026

Comic - Mario's Iconic Gun www.savestatecomic.com/comic/marios...

1182

Bryan Wyatt @brwyatt.net · 07/04/2026

So this is fun... Sent in an 8TB WD_Black SSD for warranty replacement (died after <2 months) and SanDisk sent me a DOA replacement. Did they not even QA the replacement? Or did they think I wouldn't notice? Well, we'll see how support goes with this one...

000

Bryan Wyatt @brwyatt.net · 03/04/2026

In case you've missed all the other signs you should've been asking this question... ... here's your chance to "catch up" with the rest of us. (The answer, unfortunately, is "yes, yes we are")

000

Ceph comes with a complexity (admin) and network latency and replication latency hit. But dang, I love the benefits that give you!

000

But yeah.... ~2 months old drive. At least I can just take my time and wait for the RMA, rather than having to run out to replace the drive. Redundancy is completely recovered, and I've gone from 49% usage to 52%. Honestly? Not really a big deal.

100

Technically, the Erasure-Coded bulk storage pool (3+1) had SOME risk until the placement groups on the failed drive were recovered in order to maintain redundancy... but... that's fine. Intentional trade-off for space efficiency.

100

The only real "negative impact" here would've been a few seconds of write-pause for any placement groups targeting the drive when it went offline while the cluster timed it out. Critical VM data is 4x replicated, so no risk there.

100

Seriously, though... This is definitely one of the coolest things. Lose a host or drive? Data just shuffles around on existing drives/hosts to recover redundancy. RAID with warm spares only wishes it was this cool!

100

Bryan Wyatt @brwyatt.net · 21/02/2026

So something cool happened last night. An SSD died. "Wait! That's not cool!" Except... It was just a non-event with Ceph. Data re-replicated to other drives/hosts within a few hours. ... now to get that RMA started...

100

Welp... after a month and a half, I've finally released v0.1.0 (beta) of my Distributed FFmpeg coordinator service! github.com/brwyatt/dffm... Kinda terrifying to hit this point, it's been a ton of work to get here... but... well, I believe in it... sooo... 😅

github.com

GitHub - brwyatt/dffmpeg: `dffmpeg` is a centrally-coordinated FFmpeg worker job manager

`dffmpeg` is a centrally-coordinated FFmpeg worker job manager - brwyatt/dffmpeg

000

Bryan Wyatt @brwyatt.net · 14/02/2026

There's nothing more rewarding than seeing something you've been working on for over a month finally working! Decided I didn't like rffmpeg, and decided to go overkill with a high-availability "distributed ffmpeg" coordinator service. github.com/brwyatt/dffm... Warning: consider it pre-alpha

010

Bryan Wyatt @brwyatt.net · 13/01/2026

I dunno, I just find it fascinating how hard things hit me at the time that "broke" my brain in such a way that I went from not experiencing (consciously) dreams to it being a regular experience. That's just... wild.

000

Bryan Wyatt @brwyatt.net · 13/01/2026

Like... I went from something like once every year or so as a weird occurrence... to now, it being a frequent (multiple nights per week) thing. I wish I had brain scans before, during, immediately after, and now... just to see the changes the "event" caused in my brain to make this happen.

100

Bryan Wyatt @brwyatt.net · 13/01/2026

I didn't used to dream. More accurately, I wouldn't ever recall dreaming. It just wasn't part of my typical experience. Then a particularly traumatic/painful breakup a few years ago changed that. Started as nightmares, but then developed into a more pleasant experience with regular frequency.

100

Bryan Wyatt @brwyatt.net · 07/01/2026

This just makes me chuckle. I'd rather they just not have this stupid thing popup automatically whenever I track a package, but I like to think this is tracked in someone's metrics somewhere...

000

Bryan Wyatt @brwyatt.net · 05/12/2025

Whoever at Microsoft designed the "Stay signed in?" -> [Yes] -> "Signing you OUT" workflow should be sat in the corner to think about what they did. That's even worse than the "stay signed in?" not actually keeping you signed in and asking again later anyway thing, IMO.

000

Bryan Wyatt @brwyatt.net · 27/11/2025

Sometimes, having a well-running HomeLab is less fun... Nothing is broken, and it "just works". How boring. 🥱

000

Bryan Wyatt @brwyatt.net · 17/11/2025

Or, arguably, used in defense of a state against federal invasion/over-reach... Not as a mechanism of that federal invasion/over-reach! Literally the whole point of the existence of the National Guard as a state militia in the first place!

000

Oh, and added a little helper script that just makes "mount and chroot to the system" a bit easier for things like boot maintenance/recovery. It has definitely made this round of updates/fixes a LOT easier/faster.

000

Diskless Desktop off Ceph RBD root

Can see the results here: gist.github.com/brwyatt/7984... ... or... well, at least parts of it, I guess. Some isn't necessarily there (like the netplan config).

gist.github.com

Diskless Desktop off Ceph RBD root. GitHub Gist: instantly share code, notes, and snippets.

100

DNS required DHCP (effectively, at least to be flexible). Jumbo frames was easier than I thought, just have to set it in the initramfs script AND NetPlan (or else it drops connections on boot) MAC list was annoying (no Bash arrays), but found a solution that works instead.

100