@Daniil I'm sorry for the earlier short reply. I had actually turned off my monitor, stood up, and turn away from my machine, intending to get some sleep (after having already nodded-off a couple of times), when I heard the noise for your ping. I really wasn't at the point of responding to anything other than something that needed to be handled right then.
The "probably" is because your question can't be answered definitively, without you linking to the ping you're asking about. The description "my ping about FIRE" doesn't make it unique, as there has been more than one. Thus, there's no way for me to know if I've gotten the one you're talking about. I'm wasn't upset, or anything (other than tired), nor did I intend the response to be dismissive. It's just there wasn't enough information to give an accurate answer as to seeing a specific ping.
@Daniil Since several hours ago. I brought down my EC2 instance in January, because the year-long free-trial AWS EC2 instance time had expired. While an EC2 instance (at the level that is free for the first year) can easily run an SD instance and more, it doesn't have the CPU credits to run it as the active instance 100% of the time.
Basically, at the level of the free-tier the instance is good for ~8 hours after being in standby for 24 hours, from a CPU credit POV. Obviously, you can pay for more CPU time, either as a baseline or when needed.
So, I didn't want to pay for having it available, particularly when I had a Windows box on which I could run an instance for free (currently intentionally down for other reasons). Having the EC2 instance was nice, but just not really worth the money.
@user12986714 At least for earlier generations of rpi, that's been done, but at least for the configuration used at that time, SD was a bit much for it (at that time, and may be worse now), and eventually destroyed the FLASH media, probably due to writing to the file system too much. OTOH, the available hardware options have improved. Prior to proceeding in that direction, or getting SD working on Android, I'd wanted to massively reduce the amount of writing to disk.
Even on my Windows instance, the amount of disk writing is quite noticeable. For my Windows instance, I run it off of a RAM drive, so I no longer care (although, I do care about the size of the RAM drive, which I recently had to increase), but prior to switching to using a RAM drive, it was very noticeable (drive seeking sounds) when that SD instance would become active (the SD repo was initially on a spinning drive, rather than a SSD, due to the write-volume issue).
@Daniil There hasn't really been a need at the moment. Osiris has primarily been up 98–99% of the time, and the other instances have been easily able to cover the load.
@ThomasWard np. It was definitely heading towards when I should have been asleep. I just happened to notice that both you and @iBug were around and talking. :)
@Undo Sorry. I'd intended to pre-clear things on Keybase, but I fell asleep and then @ThomasWard was immediately available when I woke up. I'm sorry for not having done so.
@Undo Thanks, but it wasn't critical enough such that I couldn't have taken the brief time to give a notification prior to making the change. I'd blame not enough caffeine, but I believe in taking responsibility for my actions, and this definitely could have been better.
@iBug When you get a chance, the GitHub password for SD has been changed, due to a potential security issue. You can get the new password on Keybase.
Potentially bad keyword in body - Positions 67-75, 376-384: Cash App, Position 229-237: cash app, Position 430-446: cash app account, Position 711-718: CashApp Potentially bad keyword in title - Position 7-14: CashApp
Can we all take a moment to recognize how Makyen is the definition of professional? Always a pleasure to work with, communicates clearly with everyone, all around great. Said it before, but we’re extremely lucky. Thanks :)
@tripleee Is Halflife Python based? Does it use ChatExchange? If so, Halflife should benefit significantly from updating to the current version of ChatExchange. ChatExchange has been adjusted to remove some artificial self-throttling of chat messages, which should provide a noticeable improvement in sending rate.
The largest improvement will be for messages that tend to be sent as a burst. For example, 4 messages sent after no activity should now take ~2 seconds, whereas they would previously have taken ChatExchange 15 seconds. For 6 messages: ~6 seconds instead of 25 seconds.
