Author: Luke-Jr 2011-09-09 06:57:03
Published on: 2011-09-09T06:57:03+00:00
The author of the post is running two bitcoinds on the same system, one as a hub node and the other as a work node. The hub node relays network activity between hundreds of nodes Eligius peers with while the work node manages mining and only connects to the hub node. However, the hub node has been dying lately without any apparent reason, leaving the work node with no connections. As a result, the work node starts responding to getworks with "not connected" messages. Interestingly, the sockets used for these messages are never cleaned up properly and eventually overflow the system's socket limits. The author is using a heavily patched 0.3.23 version, so the bug could be related to one of those patches, such as the threaded JSON-RPC. The author hasn't tried to reproduce it on an unpatched bitcoind yet, but felt like mentioning it somewhere else just in case before 0.4 final got released. Due to being extremely busy, the author might not get a chance to do so before the 0.4 release.
Updated on: 2023-06-04T19:04:19.643378+00:00