agorise/morphit
agorise/morphit
Watch
1
0
Fork
You've already forked morphit
 0
Code Issues Pull requests Projects Releases 186 Packages Wiki Activity Actions
186 releases 186 tags
  • v1.14.1 d8dcb48af8

    Morphit v1.14.1
    All checks were successful
    morphit-ci / Supply-chain audit gate (fail on new high/critical) (push) Successful in 21s
    Details
    morphit-ci / TypeScript typecheck (sweep all workspaces) (push) Successful in 46s
    Details
    morphit-ci / apps/web svelte-check (svelte-kit sync + svelte-aware tsc) (push) Successful in 37s
    Details
    morphit-ci / Integration tests (real Postgres 16) (push) Successful in 39s
    Details
    morphit-ci / ansible-lint (playbook quality gate) (push) Successful in 15s
    Details
    morphit-ci / Smoke suite (run-smokes.sh, triple-pulse) (push) Successful in 28m34s
    Details
    morphit-release / Build + publish release tarball (push) Successful in 37m17s
    Details
    Stable

    agorise released this 2026-08-30 17:58:34 +00:00 | 48 commits to main since this release

    Signed by agorise
    GPG key ID: 53524E1F1017EB9C

    Morphit v1.14.1

    A UX + hardening follow-up to v1.14.0. No new features, no database migration, no on-chain changes.

    Changed

    Fast-sync is now the default when setting up a new node. The guided install wizard (morphit-ops install) now offers fast-sync as a default-yes step: a new operator just presses Enter to restore the newest signed federation snapshot and be live in minutes, and has to actively decline to fall back to a full replay. This is what v1.14.0's release notes described; in v1.14.0 fast-sync existed only as a separate command, so this closes that gap.

    Fixed

    The stats-page RPC endpoints card now tells the truth on refresh. Clicking the card's refresh button (or reloading the page) fires a live re-probe of every node — but a node that had just been shut down kept showing its stale last-known latency instead of going offline. The card's anti-flicker rule (which correctly ignores a single jittery Tor/I2P miss on a node that's actually up) was suppressing every fresh miss, because a failed probe always reported exactly one failure. Now a missed probe is retried once, and only a genuine double-miss is reported as a real outage — so a node you shut down shows as "offline at the moment" on the next refresh, while a merely-flaky-but-alive node still doesn't false-alarm.

    fail2ban can no longer lock out your admin machine. The standard node hardening (both the Ansible role and the manual OPERATIONS.md §34 steps) now writes an ignoreip whitelist into fail2ban's jail.local, with a fail2ban_ignoreip variable for your admin IP. This prevents the failure mode where a noisy release ceremony (repeated SSH passphrase prompts → failed auths) got the operator's own IP banned mid-release.

    Notes

    • Upgrade in place; everything from v1.14.0 and earlier is included.
    Downloads