Technologie & IA
Flare rebuilds go-flare 1.15.1 after Helicon transition bug
The replacement build published on November 11 fixes a startup error on Songbird and Coston. Affected operators need to replace their installation even if it already reports version 1.15.1.

Startup failures on Songbird and Coston
Go-flare version 1.15.1 was rebuilt on November 11, 2026, after a bug was found during the transition to Helicon. According to the Flare Foundation's release notes, the initial build could prevent the C-Chain from starting on Songbird and Coston. At activation, the software found that the transition timestamp in its configuration did not match the one stored in the node's database.
The maintainers' schedule set the November 11 activation for 7 a.m. UTC on Coston, 8 a.m. on Songbird and 2 p.m. on Flare. Coston2 had crossed that milestone on November 3 at noon UTC with a prerelease. The updated notes say the November 11 bug affected the first two networks on that day's schedule. Flare and Coston2 were unaffected by this particular bug, although the maintainers recommend the corrected build for every network.
Two builds share the same version number
The Flare Foundation retained the v1.15.1 tag and moved it to revision 9eacc2d2. The release binaries and both Docker images were rebuilt from that revision. A Songbird or Coston operator who installed the first build therefore needs to download the files again and restart the node. The 1.15.1 version number alone does not distinguish the two builds.
The maintainers say the info.getNodeVersion RPC call should report avalanchego/1.15.1 and return the full corrected revision in the gitCommit field. They also say a node stopped by the bug can recover after the files are replaced and the node restarts, without a database change or a full resync. Operators building from source need to refresh their Git tags before rebuilding because the tag already in a local repository may still point to the first build.
State sync has a separate timing constraint
The release documentation also flags a constraint associated with Helicon. The C-Chain needs to produce 4,096 blocks after activation before the first state summary becomes available. During that interval, estimated at around 90 minutes in the release notes, a node started with state sync enabled may get stuck bootstrapping. The relevant setting is in the C-Chain configuration file.
On the official forum, the recovery notice asked Songbird operators to leave the setting disabled until block 137,682,944. That threshold gives operators a measurable checkpoint when the estimated time varies with block production. The released fix addresses startup for the affected nodes. The release notes give no count of nodes running the original build and no measurement of the availability of individual services during the transition.
