← 输入框|github.com|刷新
Skip to content

node: Flush the chainstate on shutdown - #117

Merged
rustaceanrob merged 2 commits into
kernel-node:masterfrom
pzafonte:flush-on-shutdown
Oct 29, 2026
Merged

rustaceanrob merged 2 commits into
kernel-node:masterfrom
pzafonte:flush-on-shutdown

Conversation

@pzafonte

Copy link
Copy Markdown
Collaborator

The kernel writes the block index and coins database when the chainstate manager is destroyed. The IPC listener thread holds a reference and is never joined. A node stopped with cli stop exits with the manager alive, and the next start syncs from genesis.

Stop the listener with a oneshot and join it after shutdown. The added test syncs a node from Core, stops both, and restarts the node on the same datadir with no peers. On master the restarted node reports genesis instead of the synced tip.

The kernel writes the block index and coins database when the
chainstate manager is destroyed. The IPC listener thread holds a
reference and is never joined. A stopped node exits with the manager
alive, and the next start syncs from genesis.

Stop the listener with a oneshot and join it after shutdown. The
oneshot needs tokio's sync feature.
Every test uses a fresh datadir, and none checks what a stopped node
leaves on disk. Sync a node from Core, stop both, and restart the node
with no peers. The tip it reports can only come from disk.

stop returns the datadir for reuse. start_in removes the stale socket,
which would otherwise pass the readiness check early.
@rustaceanrob
rustaceanrob merged commit d421e33 into kernel-node:master Oct 29, 2026
2 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants