rename: SYS_WATCHDOG_FAILED_FILE → DOCKER_WATCHDOG_FAILED_FILE
Skip list belongs to docker_watchdog, not stability_watchdog (legacy naming from when docker+system were the only two watchdogs). Rename variable and physical file system_watchdog_failed.db → docker_watchdog_failed.db across all files. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
This commit is contained in:
co-authored by
Claude Sonnet 4.6
parent
9d6772fbd6
commit
070f9caf7f
@@ -134,7 +134,7 @@
|
||||
# ==============================================================================================
|
||||
#
|
||||
# WATCHDOG_STATE_FILE — strike counts, daemon flags (STATE_DIR — survives reboots)
|
||||
# SYS_WATCHDOG_FAILED_FILE — container skip list (STATE_DIR — survives reboots)
|
||||
# DOCKER_WATCHDOG_FAILED_FILE — container skip list (STATE_DIR — survives reboots)
|
||||
# WATCHDOG_CONTAINER_RESTART_LOG — restart history for loop detection (DATA_DIR)
|
||||
# RW_STATE_FILE — read-only: resource_watchdog RAM emergency flag
|
||||
#
|
||||
@@ -255,7 +255,7 @@ validate_unraid_cmd "/usr/local/emhttp/plugins/dynamix/scripts/notify" "
|
||||
|
||||
# Ensure state files exist
|
||||
touch "$WATCHDOG_STATE_FILE" "$WATCHDOG_CONTAINER_RESTART_LOG" \
|
||||
"$SYS_WATCHDOG_FAILED_FILE" 2>/dev/null
|
||||
"$DOCKER_WATCHDOG_FAILED_FILE" 2>/dev/null
|
||||
|
||||
# Timeout for all docker commands — prevents hung daemon from stalling the watchdog
|
||||
DOCKER_TIMEOUT=15
|
||||
@@ -306,14 +306,14 @@ set_strikes() {
|
||||
|
||||
# Check if container is on the persistent skip list
|
||||
is_skipped() {
|
||||
grep -q "^${1}$" "$SYS_WATCHDOG_FAILED_FILE" 2>/dev/null
|
||||
grep -q "^${1}$" "$DOCKER_WATCHDOG_FAILED_FILE" 2>/dev/null
|
||||
}
|
||||
|
||||
# Add container to persistent skip list — manual intervention required to recover
|
||||
add_to_skip_list() {
|
||||
local container="$1" reason="$2"
|
||||
if ! is_skipped "$container"; then
|
||||
echo "$container" >> "$SYS_WATCHDOG_FAILED_FILE"
|
||||
echo "$container" >> "$DOCKER_WATCHDOG_FAILED_FILE"
|
||||
error "$container added to skip list — $reason"
|
||||
queue_notify "$container added to skip list on $(hostname) — $reason — manual intervention needed" "critical"
|
||||
fi
|
||||
@@ -321,7 +321,7 @@ add_to_skip_list() {
|
||||
|
||||
# Remove container from skip list — called when container is seen running again
|
||||
remove_from_skip_list() {
|
||||
sed -i "/^${1}$/d" "$SYS_WATCHDOG_FAILED_FILE" 2>/dev/null
|
||||
sed -i "/^${1}$/d" "$DOCKER_WATCHDOG_FAILED_FILE" 2>/dev/null
|
||||
warn "$1 recovered — removed from skip list ✅"
|
||||
}
|
||||
|
||||
|
||||
Reference in New Issue
Block a user