cardano-node 11.1.3 reverts an IPv4 byte-order change made in 11.1.2
Serguey Asael Shinder1 min readIntersectMBO published cardano-node 11.1.3 on 29 September. The release has a single fix, in how the ledger decodes IPv4 addresses. Releases before the 11.1 line decoded them in network byte order; version 11.1.2 inadvertently changed that, and 11.1.3 reverts it to network order.
IPv4 addresses appear in the ledger in stake pool registrations, where pools publish the relays other nodes connect to. A byte-order error there does not corrupt funds, but it changes how an address written by one node version is read by another, which is exactly the sort of discrepancy a network of mixed versions cannot afford.
The known issues carry over from 11.1.1. Memory use while syncing is slightly higher than on 11.0.1, though far below the regression in 11.1.0, and memory on tip rises a little when the node takes ledger snapshots. The notes list one breaking change, in plutus-ledger-api: plc optimise no longer accepts the --certify and --certifier- options. Benchmarks and system tests from 11.1.1 apply unchanged, according to the notes.
Minimum requirements stay at two cores and 24 GB of RAM with the in-memory ledger backend, or 8 GB with the on-disk backend, still marked as pending confirmation, plus 300 GB of storage. Block production and relays remain supported on Linux only.
Why it matters
11.1.2 shipped on 17 September to trim memory in time lock scripts; twelve days later the next patch undoes a side effect in a part of the code the release was not about. Operators who moved to 11.1.2 should take 11.1.3, and the episode is a reminder that serialisation details are part of the protocol whether or not a release mentions them.