You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
There's been an issue for as long as I can remember w/ Comet where peers often disconnect from each other despite all peers operating using unmodified software. For example, in testground we see a large number of peer disconnections despite all nodes in the experiment running stock software. Its worth noting that the default timeouts are being used (45 seconds for receiving a pong)
Here are all of the different errors that resulted in a peer getting disconnected from a single testground
experiment
find the cause of most if not all of the above issues and fix them. This would likely involve re-running these traces in knuu considering that is the future of our testing infra, and then addressing them. Considering all nodes in the network tests are behaving honestly and with adequate resources, in theory they should not disconnect.
The text was updated successfully, but these errors were encountered:
It could be the case that nodes are being overwhelmed, in which case debugging bottlenecks in the p2p stack would also address the issues here. Therefore, it makes sense to debug that before moving forward with this.
There's been an issue for as long as I can remember w/ Comet where peers often disconnect from each other despite all peers operating using unmodified software. For example, in testground we see a large number of peer disconnections despite all nodes in the experiment running stock software. Its worth noting that the default timeouts are being used (45 seconds for receiving a pong)
Here are all of the different errors that resulted in a peer getting disconnected from a single testground experiment
AC
find the cause of most if not all of the above issues and fix them. This would likely involve re-running these traces in knuu considering that is the future of our testing infra, and then addressing them. Considering all nodes in the network tests are behaving honestly and with adequate resources, in theory they should not disconnect.
The text was updated successfully, but these errors were encountered: