de1a274f6a
- clear-all-transaction-pools: use PROXMOX_SSH_USER (never root@pam for SSH); align R630 host with ip-addresses PROXMOX_R630_01; document post-clear RPC delay - preflight-chain138-deploy: CHAIN138_PREFLIGHT_RPC_URL for nonce/RPC checks when Core 2101 is restarting (e.g. after pool clear) but public RPC is up Made-with: Cursor
143 lines
6.2 KiB
Bash
Executable File
143 lines
6.2 KiB
Bash
Executable File
#!/usr/bin/env bash
|
|
# Clear transaction pools on all Besu nodes (RPC and Validators)
|
|
# This script clears transaction pool databases to remove stuck transactions
|
|
#
|
|
# SSH: uses PROXMOX_SSH_USER from config/ip-addresses.conf (root). If .env sets PROXMOX_USER=root@pam for the API,
|
|
# that value is NOT used for SSH (see PROXMOX_USER= assignment below).
|
|
# Afterward: Core RPC (2101) can take several minutes to bind JSON-RPC while RocksDB opens/compacts; public RPC may recover first.
|
|
|
|
set -euo pipefail
|
|
|
|
# Load IP configuration
|
|
SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
|
|
PROJECT_ROOT="$(cd "$SCRIPT_DIR/.." && pwd)"
|
|
source "${PROJECT_ROOT}/config/ip-addresses.conf" 2>/dev/null || true
|
|
|
|
# Shell SSH must be root@host — not root@pam@host (.env often sets PROXMOX_USER=root@pam for API).
|
|
PROXMOX_SSH_USER="${PROXMOX_SSH_USER:-root}"
|
|
[[ "$PROXMOX_SSH_USER" == *"@"* ]] && PROXMOX_SSH_USER="root"
|
|
PROXMOX_USER="$PROXMOX_SSH_USER"
|
|
PROXMOX_ML110="${PROXMOX_ML110:-${PROXMOX_HOST_ML110:-192.168.11.10}}"
|
|
PROXMOX_R630="${PROXMOX_R630:-${PROXMOX_R630_01:-${PROXMOX_HOST_R630_01:-192.168.11.11}}}"
|
|
|
|
# Colors
|
|
RED='\033[0;31m'
|
|
GREEN='\033[0;32m'
|
|
YELLOW='\033[1;33m'
|
|
BLUE='\033[0;34m'
|
|
CYAN='\033[0;36m'
|
|
NC='\033[0m'
|
|
|
|
log_info() { echo -e "${BLUE}[INFO]${NC} $1"; }
|
|
log_success() { echo -e "${GREEN}[✓]${NC} $1"; }
|
|
log_warn() { echo -e "${YELLOW}[⚠]${NC} $1"; }
|
|
log_error() { echo -e "${RED}[✗]${NC} $1"; }
|
|
log_section() { echo -e "\n${CYAN}━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━${NC}"; echo -e "${CYAN}$1${NC}"; echo -e "${CYAN}━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━${NC}\n"; }
|
|
|
|
echo "=== Clear Transaction Pools on All Nodes ==="
|
|
echo ""
|
|
|
|
# Function to clear transaction pool for a node
|
|
clear_node_pool() {
|
|
local VMID=$1
|
|
local HOST=$2
|
|
local NODE_TYPE=$3
|
|
local SSH_TARGET="${PROXMOX_USER}@${HOST}"
|
|
|
|
log_info "Clearing transaction pool for $NODE_TYPE (VMID $VMID on $HOST)..."
|
|
|
|
# Stop the service
|
|
log_info " Stopping service..."
|
|
ssh -o ConnectTimeout=10 -o StrictHostKeyChecking=no "$SSH_TARGET" \
|
|
"pct exec $VMID -- systemctl stop besu-validator 2>/dev/null || pct exec $VMID -- systemctl stop besu-rpc-core 2>/dev/null || pct exec $VMID -- systemctl stop besu-rpc.service 2>/dev/null || pct exec $VMID -- systemctl stop besu-rpc 2>/dev/null || true" 2>&1 | grep -v "Configuration file" || true
|
|
|
|
sleep 2
|
|
|
|
# Find and clear transaction pool database
|
|
log_info " Clearing transaction pool database..."
|
|
CLEAR_RESULT=$(ssh -o ConnectTimeout=10 -o StrictHostKeyChecking=no "$SSH_TARGET" \
|
|
"pct exec $VMID -- bash -c '
|
|
DATA_DIRS=\"/data/besu /var/lib/besu\"
|
|
for DATA_DIR in \$DATA_DIRS; do
|
|
if [ -d \"\$DATA_DIR\" ]; then
|
|
# Find transaction pool database files
|
|
find \"\$DATA_DIR\" -type d -name \"*pool*\" -exec rm -rf {} \; 2>/dev/null || true
|
|
find \"\$DATA_DIR\" -type f -name \"*transaction*\" -delete 2>/dev/null || true
|
|
find \"\$DATA_DIR\" -type f -name \"*txpool*\" -delete 2>/dev/null || true
|
|
echo \"Cleared: \$DATA_DIR\"
|
|
fi
|
|
done
|
|
'" 2>&1 | grep -v "Configuration file" || echo "Cleared")
|
|
|
|
if [ -n "$CLEAR_RESULT" ]; then
|
|
log_success " Transaction pool cleared"
|
|
else
|
|
log_warn " Could not clear transaction pool (may not exist)"
|
|
fi
|
|
|
|
# Restart the service
|
|
log_info " Restarting service..."
|
|
ssh -o ConnectTimeout=10 -o StrictHostKeyChecking=no "$SSH_TARGET" \
|
|
"pct exec $VMID -- systemctl start besu-validator 2>/dev/null || pct exec $VMID -- systemctl start besu-rpc-core 2>/dev/null || pct exec $VMID -- systemctl start besu-rpc.service 2>/dev/null || pct exec $VMID -- systemctl start besu-rpc 2>/dev/null || true" 2>&1 | grep -v "Configuration file" || true
|
|
|
|
sleep 3
|
|
|
|
# Verify service is running
|
|
STATUS=$(ssh -o ConnectTimeout=5 -o StrictHostKeyChecking=no "$SSH_TARGET" \
|
|
"pct exec $VMID -- systemctl is-active besu-validator 2>/dev/null || pct exec $VMID -- systemctl is-active besu-rpc-core 2>/dev/null || pct exec $VMID -- systemctl is-active besu-rpc.service 2>/dev/null || pct exec $VMID -- systemctl is-active besu-rpc 2>/dev/null || echo 'unknown'" 2>&1 | grep -v "Configuration file" || echo "unknown")
|
|
|
|
if [ "$STATUS" = "active" ]; then
|
|
log_success " Service restarted and active"
|
|
else
|
|
log_warn " Service status: $STATUS"
|
|
fi
|
|
|
|
echo ""
|
|
}
|
|
|
|
# Clear validators
|
|
log_section "Clearing Validator Transaction Pools"
|
|
|
|
VALIDATORS=(
|
|
"1000:$PROXMOX_R630:Validator"
|
|
"1001:$PROXMOX_R630:Validator"
|
|
"1002:$PROXMOX_R630:Validator"
|
|
"1003:$PROXMOX_ML110:Validator"
|
|
"1004:$PROXMOX_ML110:Validator"
|
|
)
|
|
|
|
for validator in "${VALIDATORS[@]}"; do
|
|
IFS=':' read -r VMID HOST TYPE <<< "$validator"
|
|
clear_node_pool "$VMID" "$HOST" "$TYPE"
|
|
done
|
|
|
|
# Clear RPC Core (2101)
|
|
log_section "Clearing RPC Transaction Pool (2101)"
|
|
|
|
if ssh -o ConnectTimeout=5 -o StrictHostKeyChecking=no "${PROXMOX_USER}@${PROXMOX_ML110}" \
|
|
"pct list | grep -q '2101'" 2>/dev/null; then
|
|
clear_node_pool 2101 "$PROXMOX_ML110" "RPC"
|
|
elif ssh -o ConnectTimeout=5 -o StrictHostKeyChecking=no "${PROXMOX_USER}@${PROXMOX_R630}" \
|
|
"pct list | grep -q '2101'" 2>/dev/null; then
|
|
clear_node_pool 2101 "$PROXMOX_R630" "RPC"
|
|
else
|
|
log_warn "RPC node (2101) not found on either host"
|
|
fi
|
|
|
|
# Clear RPC Public (2201) — often used when Core is down; ensures deploy txs not stuck
|
|
log_section "Clearing RPC Public (2201)"
|
|
R630_02="${PROXMOX_R630_02:-${PROXMOX_HOST_R630_02:-192.168.11.12}}"
|
|
if ssh -o ConnectTimeout=5 -o StrictHostKeyChecking=no "${PROXMOX_USER}@${R630_02}" \
|
|
"pct list | grep -q '2201'" 2>/dev/null; then
|
|
clear_node_pool 2201 "$R630_02" "RPC Public"
|
|
else
|
|
log_warn "RPC Public (2201) not found on ${R630_02}"
|
|
fi
|
|
|
|
log_section "Transaction Pool Clear Complete"
|
|
|
|
echo "Next steps:"
|
|
echo " 1. Wait 30-60 seconds for nodes to fully restart"
|
|
echo " 2. Check pending transactions: bash scripts/check-pending-transactions.sh"
|
|
echo " 3. Monitor health: bash scripts/monitoring/monitor-blockchain-health.sh"
|