Conversation

Notices

1.  (mint@ryona.agency)'s status on Wednesday, 01-Nov-2023 18:53:47 JST 
   UPS went off. See you in N hours.
   • † top dog :pedomustdie: likes this.
   • † top dog :pedomustdie: (dcc@annihilation.social)'s status on Wednesday, 01-Nov-2023 18:54:39 JST † top dog :pedomustdie:

     in reply to
     @mint Have a fun time
      likes this.
   •  (mint@ryona.agency)'s status on Wednesday, 01-Nov-2023 20:05:54 JST 

     in reply to
     We're back.
     Machismo and † top dog :pedomustdie: like this.
   •  (mint@ryona.agency)'s status on Wednesday, 01-Nov-2023 20:07:38 JST 

     in reply to
     If I had to choose between ten five minute outages every day and one two hour outage every few weeks, I'd choose the former. At least it can be mitigated.
     Machismo likes this.
   •  (mint@ryona.agency)'s status on Wednesday, 01-Nov-2023 20:31:38 JST 

     in reply to
     [ ID] Interval Transfer Bitrate Retr [ 5] 0.00-10.00 sec 452 KBytes 370 Kbits/sec 41 sender [ 5] 0.00-10.14 sec 401 KBytes 324 Kbits/sec receiver

     You've got to be shitting me.

   •  (mint@ryona.agency)'s status on Wednesday, 01-Nov-2023 20:32:48 JST 

     in reply to

     Switched wireguard back to ipv6, much better but still not great.

     [ ID] Interval Transfer Bitrate Retr [ 5] 0.00-10.00 sec 36.0 MBytes 30.2 Mbits/sec 16 sender [ 5] 0.00-10.13 sec 32.6 MBytes 27.0 Mbits/sec receiver
   • Shrek (sarvo@novoa.nagoya)'s status on Wednesday, 01-Nov-2023 21:14:30 JST Shrek

     in reply to

     @mint@ryona.agency I remember wireguard giving me some strange bottlenecks, I don't remember how I fixed it tho

      likes this.
   •  (mint@ryona.agency)'s status on Wednesday, 01-Nov-2023 23:36:41 JST 

     in reply to
     Literally no difference between wireguard over v4 and v6, don't even remember which is which.
     Husky_1698849319140_JLTGHPKQLA.…
     Husky_1698849322144_XZGZ1SDUEQ.…
   •  (mint@ryona.agency)'s status on Wednesday, 01-Nov-2023 23:39:30 JST 

     in reply to
     Maybe it's getting throttled statewide, I'll try wrapping it in udp2raw later.
   •  (mint@ryona.agency)'s status on Thursday, 02-Nov-2023 00:05:30 JST 

     in reply to
     Went to the server to download udp2raw on it, and the download speed from Frantech to Github is 105KB/s. Doesn't seem right. No abnormalities in I/O or CPU usage, nothing in dmesg indicating the problem on my side. Perhaps the whole dataventer is getting DDoS'd or something.
   •  (mint@ryona.agency)'s status on Thursday, 02-Nov-2023 00:10:23 JST 

     in reply to
     Maybe it's upstream.
     Husky_1698851360700_G2755PL9SH.…
     Husky_1698851376271_XHCQTSE9PW.…
   • pwm (pwm@crlf.ninja)'s status on Thursday, 02-Nov-2023 00:16:05 JST pwm

     in reply to
     @mint the exfil rate on my seedbox with frantech has been abysmal. Lucky to push 2-3 MiB/s consistently, with frequent drops. My internet at the casa is about 10x that so I know it's not me unless it's fuckery with my ISP to the DC. Not that it's urgent or anything but I would really enjoy it being more like I would expect
      likes this.
   •  (mint@ryona.agency)'s status on Thursday, 02-Nov-2023 01:52:57 JST 

     in reply to
     :wereback:
     Screenshot_20231101_195217.png
   •  (mint@ryona.agency)'s status on Thursday, 02-Nov-2023 01:54:11 JST 

     in reply to
     Yggdrasil is still kinda slow, I guess more people are updating to 0.5 and it overloads the existing peers.
     † top dog :pedomustdie: likes this.
   •  (mint@ryona.agency)'s status on Thursday, 02-Nov-2023 02:07:34 JST 

     in reply to

     • pwm
     @pwm Check if it's better now, seems like whatever issue it was has been resolved as now I'm getting proper triple-digit Mbps.
   • pwm (pwm@crlf.ninja)'s status on Thursday, 02-Nov-2023 02:12:49 JST pwm

     in reply to
     @mint I am not, it seems to be fluctuating wildly.
      likes this.
   • pwm (pwm@crlf.ninja)'s status on Thursday, 02-Nov-2023 02:13:25 JST pwm

     in reply to

     • pwm
     @mint it could just be lossy, I haven't checked specifically
      likes this.
   • pwm (pwm@crlf.ninja)'s status on Thursday, 02-Nov-2023 02:23:37 JST pwm

     in reply to

     • pwm
     @mint not lossy, feels like cpu is too busy, giant std dev and max ping times
      likes this.
   • 7666 (7666@comp.lain.la)'s status on Thursday, 02-Nov-2023 02:52:28 JST 7666

     in reply to

     • pwm
     @pwm @mint frantech box? I have had one of my four edge nodes eating ping times of 2000ms or greater to the *default gateway*. After analyzing top, the spikes correlates to massive steal spikes, likely NOT caused by a customer VM because the whole bloody box hangs and customers can't really do that without having every single core allocated to them.
     
     I threw in a ticket and the problem somewhat fixed itself after some customers got capped but the issue still didn't quite go away until last night. Will analyze the data again to see.
      likes this.
   • pwm (pwm@crlf.ninja)'s status on Thursday, 02-Nov-2023 02:53:33 JST pwm

     in reply to

     • 7666
     @7666 @mint Yeah frantech. It's funny because my edge for crlf does not exhibit this behavior, it is in the same DC, but presumably a different physical box. I'm not getting 2k ms pings to default gateway but I will have latency spikes up to 200 ms from a normal 40-60ms to my home server over a wg tunnel.
      likes this.
   • 7666 (7666@comp.lain.la)'s status on Thursday, 02-Nov-2023 14:53:06 JST 7666

     in reply to

     • pwm
     @pwm @mint here's what it looks like today, and that's after some jackasses were capped. I'd still consider this generally an issue.
      likes this.