DNS seeds returning gone peers [combined summary]

• DNS seeds returning gone peers Matt Corallo 2011-08-03 15:32:00+00:00
• DNS seeds returning gone peers Gregory Maxwell 2011-08-03 14:48:33+00:00
• DNS seeds returning gone peers Mike Hearn 2011-08-03 14:39:57+00:00
• DNS seeds returning gone peers Christian Decker 2011-08-03 14:37:12+00:00
• DNS seeds returning gone peers Douglas Huff 2011-08-03 14:28:59+00:00
• DNS seeds returning gone peers Rick Wesson 2011-08-03 14:18:26+00:00
• DNS seeds returning gone peers Mike Hearn 2011-08-03 14:10:56+00:00
• DNS seeds returning gone peers Rick Wesson 2011-08-03 14:00:24+00:00
• DNS seeds returning gone peers Mike Hearn 2011-08-03 12:40:48+00:00
• DNS seeds returning gone peers Caleb James DeLisle 2011-08-03 12:27:37+00:00
• DNS seeds returning gone peers Mike Hearn 2011-08-03 12:17:48+00:00
• DNS seeds returning gone peers Johannes Henninger 2011-08-03 12:15:07+00:00
• DNS seeds returning gone peers Mike Hearn 2011-08-03 12:00:51+00:00
• DNS seeds returning gone peers Matt Corallo 2011-08-03 11:38:19+00:00
• DNS seeds returning gone peers Mike Hearn 2011-08-03 10:04:11+00:00

Published on: 2011-08-03T15:32:00+00:00

Summary:

Bitcoin core developer Gregory Maxwell and Mike Hearn discuss a problem with the peer-to-peer network being starved for connectable slots. Hearn suggests increasing the anti-DoS limits in version 0.4, but Maxwell argues that it could cause nodes on lower memory systems to run out of memory. The issue persists, and they need to find a solution.In another conversation, Hearn suggests a custom DNS server to resolve the issue of bad nodes appearing in DNS seeds, which is slowing down peer bringup for Android apps. Rick Wesson offers to contribute to the DNS seeding project. They discuss sorting nodes by version and other factors and suggest using Matt's code in PHP or developing a Java app with BitCoinJ+sqlite libraries.Hearn posts on the Bitcoin-development mailing list about the DNS seed issue and hopes his friend will take on DNS seeding as a project. Wesson expresses interest in contributing and asks for clarification on long-lived peers. Hearn suggests starting from Matt's code or writing a Java app linked with BitCoinJ+sqlite.Hearn discusses the issue of bad nodes appearing in DNS seeds in an email thread. He mentions that only 19 out of 48 IPs returned are usable, affecting peer bringup for Android apps. He suggests a custom DNS server to find long-lived peers running the latest version. Corallo points to existing code on GitHub and suggests using a different DNS server that pulls directly from the database.The discussion also touches on DNS caching by ISPs and the use of randomizing DNS requests. It is suggested to do a lookup on a random cookie as a subdomain.The conversation highlights the need for a solution to the limited number of slots and the presence of bad nodes in DNS seeds. Suggestions include increasing anti-DoS limits, developing a custom DNS server, and sorting nodes by version. There is interest in contributing to the DNS seeding project and utilizing existing code. The issue of DNS caching by ISPs is also mentioned.

Updated on: 2023-08-01T02:12:32.215258+00:00