I had hoped to go out to a show tonight, but nope... Taking a drive at dusk to pick something up at the store for my mom used up all of the mental energy I had, and by the time I got home, I was in full blown headache mode.
One of the things I hate about winter is how early it gets dark. Being out at dusk is SO much more overstimulating in the winter because other people are still out and about when it gets dark at 4, so you get a face full of headlights when you're driving around.... and the mixture of astigmatism making lights looking streaky and autism's overstimulation.... I just can't.
So, once again I've cancelled my plans to stay home in the dark because it's too much to take a drive and go out tonight, and I am so frustrated about it and my lack of planning to make it work.
@masukomi@Susan60@spika@crashglasshouses@actuallyautistic Yeah, add that lots of newer cars have overly bright headlights and/or poorly aimed/designed ones that blind you like high beams...I can relate. Though I do also have a astigmatism.....I'm sure that doesn't help.
Just had a big run of what looks like spam signups to Infosec.exchange. They’re suspended now. I can’t guarantee some weren’t legit, so apologies if I suspended your sock account.
@kurtseifried @joshbressers@jerry@Viss Agreed it's a nice feature. The downside is making sure you attach mount points or volumes for persistent data. There are some apps that are or still install in /opt. Like Apache Nifi, and sometimes Tomcat....
Fedia.io remains problematic. I just found that my cron job that cleans up broken linkages between tables in the postgres database stopped working and so that's why the error 500's are back. I will get that fixed, but I have to tell you that I am wondering about the sensibility in continuing to run kbin on that instance. It's held together with tape and gum.
Now that Hetzner has 64core/128 thread sapphire rapids servers available AND they are not stupid expensive, I think I am going to start planning a migration to those running on 10gbps networks.
The number of active accounts is on a decline though, so a bit hesitant to pull the trigger.
Ok. I think I had a breakthrough on the nginx load balancer issue. The problem is a bit complicated, but the issue turned out to be an inconsistent configuration between the two app servers. At one point, I had to add a statement to the mastodon config that basically extends the timeout for s3 storage transactions because backblaze is soooo sloooowwww. The new node had a version of the config that didn’t contain this time-out modification.
I think it may be time to move to Wasabi. I have workaround stacked on top of workaround to accommodate backblaze limitations. Or maybe I move back to minio now that we have a CDN in front.
@jerry@stevesplace Docker has its pros and cons. One of its advantages when in a orchestration like Kubernetes is being able to roll out a HA deployment with somewhat greater ease....
Being an admin of a large-ish instance has been quite an eye opening experience into how awful some people are towards trans people. I really don’t understand the energy they put into hating people.