Closed
Description
After a couple of days of a workspace being online, it may sometimes become unreachable. When trying a coder ping -v against it, the following is observed in the logs:
2023-07-21 15:04:52.141 [debu] wgengine: ping(fd7a:115c:a1e0:49d6:b259:b7ac:b1b2:48f4): no matching peer
2023-07-21 15:04:52.203 [debu] wgengine: ping(fd7a:115c:a1e0:49d6:b259:b7ac:b1b2:48f4): no matching peer
2023-07-21 15:04:52.282 [debu] wgengine: ping(fd7a:115c:a1e0:49d6:b259:b7ac:b1b2:48f4): no matching peer
2023-07-21 15:04:52.403 [debu] wgengine: ping(fd7a:115c:a1e0:49d6:b259:b7ac:b1b2:48f4): no matching peer
2023-07-21 15:04:52.541 [debu] wgengine: ping(fd7a:115c:a1e0:49d6:b259:b7ac:b1b2:48f4): no matching peer
2023-07-21 15:04:52.850 [debu] wgengine: ping(fd7a:115c:a1e0:49d6:b259:b7ac:b1b2:48f4): no matching peer
2023-07-21 15:04:53.547 [debu] wgengine: ping(fd7a:115c:a1e0:49d6:b259:b7ac:b1b2:48f4): no matching peer
...
It seems like something is preventing the client from receiving the agent node when running the new PG coordinator in HA mode.