From f77eb81be082a3b2b1af93670a9194080b7cac8d Mon Sep 17 00:00:00 2001
From: bussyjd <bussyjd@users.noreply.github.com>
Date: Tue, 14 Jul 2026 15:26:58 +0400
Subject: [PATCH] fix(network): stop reth liveness probe from killing
 consistency recovery
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

The upstream ethereum-node chart's reth subchart defaults to a tcp
liveness probe on p2p (30303) with failureThreshold 3 (~12 min grace).
Reth keeps that port closed while running post-crash check_consistency
recovery, which can take far longer on a large datadir — so the kubelet
SIGKILLs it mid-recovery, each kill adds more drift to heal on the next
start, and the node crash-loops permanently (observed: 637 restarts,
~129k blocks of StoragesHistory drift).

Override the probe with failureThreshold 10000 so it effectively never
fires; a genuinely dead reth exits the container on its own. Scoped to
reth only — geth/nethermind/besu probe http-rpc and erigon probes
metrics, none of which close during recovery.

Claude-Session: https://claude.ai/code/session_01VLQSsnH9WdAsnVYGTd2omr
---
 .../embed/networks/ethereum/helmfile.yaml.gotmpl   | 14 ++++++++++++++
 1 file changed, 14 insertions(+)

diff --git a/internal/embed/networks/ethereum/helmfile.yaml.gotmpl b/internal/embed/networks/ethereum/helmfile.yaml.gotmpl
index 048f8388..24422bc2 100644
--- a/internal/embed/networks/ethereum/helmfile.yaml.gotmpl
+++ b/internal/embed/networks/ethereum/helmfile.yaml.gotmpl
@@ -105,6 +105,20 @@ releases:
             # renovate: datasource=github-releases depName=erigontech/erigon
             tag: v3.5.1
             {{- end }}
+          {{- if eq .Values.executionClient "reth" }}
+          # reth keeps p2p (30303) closed while running post-crash consistency
+          # recovery, which can exceed any bounded probe window on a large
+          # datadir. The chart's default tcp liveness probe then SIGKILLs it
+          # mid-recovery, forcing a longer recovery on the next start — a
+          # permanent crash loop. A genuinely dead process exits the container
+          # on its own, so make the probe effectively never fire.
+          livenessProbe:
+            tcpSocket:
+              port: p2p-tcp
+            initialDelaySeconds: 60
+            periodSeconds: 120
+            failureThreshold: 10000
+          {{- end }}
           persistence:
             enabled: true
             size: {{ if (index .Values "executionStorageSize" | default "") }}{{ index .Values "executionStorageSize" }}{{ else }}{{ if eq .Values.network "mainnet" }}{{ if eq .Values.mode "archive" }}4500Gi{{ else }}500Gi{{ end }}{{ else }}{{ if eq .Values.mode "archive" }}300Gi{{ else }}100Gi{{ end }}{{ end }}{{ end }}