| onefang | mathew: A few cron scripts do that random sleep thing, so that not everything cron has to run is running all at the same time. Usually it doesn't matter, things get done and your computer isn't bogged down regularly. | 00:06 |
|---|---|---|
| rwp | mathew, Let's say there are a million systems all running Devuan and all of them have a cron that runs apt update. What do you think would happen to the *servers* if a million Devuan systems all decided to update from them all on the same cron without any random sleeps? | 02:38 |
| rwp | It would be the same as a distributed denial of service attack. It's called the "Thundering Herd" problem. | 02:39 |
| rwp | So there is a random sleep added so that the million systems scattered about won't all run at the same time. It's done by design. | 02:39 |
| FatPhil | RSS/Atom broke that, back in the day. on the hour, every hour, lots of webservers were getting hammered. | 02:49 |
| rwp | Release the Thundering Herd! | 03:24 |
| rrq | I thought that logic was introduced as a attempt to ensure the client succeeding their request without having to program a more robust re-try... | 03:24 |
| FatPhil | I always set my scraping/updating cronjobs at 37 minutes past the hour, because that's the most random number. | 03:37 |
| * FatPhil puts on his asbestos underpants | 03:37 | |
| cousin_luigi | Having trouble with pppoe-relay and start-stop-daemon. It doesn't create a pidfile of its own, so I added --make-pidfile and later --background, but the PID doesn't match the process. | 09:00 |
| cousin_luigi | What do I do? | 09:00 |
| CueXXIII | cousin_luigi: if you use --background in start-stop-daemon you might want to tell pppoe-relay not to background itself again, i.e. call it with -F | 09:03 |
| cousin_luigi | CueXXIII: Oh, right! | 09:05 |
| cousin_luigi | CueXXIII: That worked perfectly, thanks. | 09:24 |
| jirib | i'm really curious if there is a GDM bug for Devuan; with default stable install it does not load | 09:52 |
| h3at | why does tasksel | 10:06 |
| h3at | just update | 10:06 |
| h3at | but installer updated at 2016 | 10:06 |
| jirib | since the installer was connected to the internet, it got updates | 10:07 |
| h3at | it update 2 months ago, but not shown on git log | 10:11 |
| sfox | updoots | 10:16 |
| h3at | updeens | 10:22 |
| data41201 | q | 10:30 |
| sfox | p | 10:39 |
| h3at | one night in nyc | 10:41 |
| gnarface | jirib: how are you trying to start it? | 11:02 |
| gnarface | just by the stock init script? | 11:02 |
| gnarface | or manually? | 11:02 |
| gnarface | could be missing graphics drivers or permissions | 11:05 |
| gnarface | i'd check the error log | 11:06 |
| jirib | gnarface: default is sysv init; yes it has still some issue with perm/seats... i was generally curious if this is widely known issue... | 11:07 |
| gnarface | dunno | 11:08 |
| gnarface | starting as your regular user, you'd need to be in the video group... starting from the init script though i wouldn't expect a problem | 11:09 |
| gnarface | there might be a seatd error that's a red herring, you don't actually need that | 11:10 |
| gnarface | you might need your actual video card's non-free drivers though if gdm expects working hardware accel... i don't know if it does but i recall that slim has in the past (though maybe that was a bug) | 11:10 |
| jirib | slim works fine | 11:51 |
| jirib | lighdm also works | 11:51 |
| gnarface | jirib: any other errors in the Xorg log? | 13:54 |
| jirib | gnarface: i'm sorry i'm busy now with work which needs to be done to pay my bills ;) i appreciate your interest and will try to report a bug later | 15:44 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!