From 128175509342cd31f372140b8558dba292ef047f Mon Sep 17 00:00:00 2001
From: Leandro Motta Barros <leandro@balena.io>
Date: Wed, 17 Apr 2024 19:57:59 -0300
Subject: [PATCH] Work around uuid file corruption in balenaEngine health check

We have detected one more way in which the uuid file used by containerd
can get corrupted. This time, the file is not empty, but doesn't contain
a valid UUID either.

This commit thus extends the existing workaround to also handle this
case.

See https://github.com/balena-os/balena-engine/issues/322

Signed-off-by: Leandro Motta Barros <leandro@balena.io>
Change-type: patch
---
 .../balena/balena/balena-healthcheck          | 38 +++++++++++++++----
 1 file changed, 30 insertions(+), 8 deletions(-)

diff --git a/meta-balena-common/recipes-containers/balena/balena/balena-healthcheck b/meta-balena-common/recipes-containers/balena/balena/balena-healthcheck
index ebd65a94a0..faf24e7e34 100644
--- a/meta-balena-common/recipes-containers/balena/balena/balena-healthcheck
+++ b/meta-balena-common/recipes-containers/balena/balena/balena-healthcheck
@@ -7,24 +7,46 @@ set -o errexit
 BALENAD_SOCKET="/run/balena.sock"
 CONTAINERD_SOCKET="/run/balena-engine/containerd/balena-engine-containerd.sock"
 
+# Checks if the file $1 contains an UUID in any of the formats described here:
+# https://github.com/google/uuid/blob/6e10cd1027e225e3ad7bfcc13c896abd165b02ef/uuid.go#L189
+contains_valid_uuid()
+{
+    # This matches xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx, which is the one format
+    # we have seen in practice, and is also the base for most other supported
+    # formats.
+    BASE_UUID_REGEX="[[:xdigit:]]{8}-([[:xdigit:]]{4}-){3}[[:xdigit:]]{12}"
+
+    # Check for all the possible formats, starting with the most common one, to
+    # benefit from the short-circuited evaluation. The contents of the uuid file
+    # must match these formats exactly, so we anchor the regexes both at the
+    # beginning (^) and the end ($).
+    grep -E          "^${BASE_UUID_REGEX}$"   "$1" > /dev/null || \
+    grep -E        "^\{${BASE_UUID_REGEX}\}$" "$1" > /dev/null || \
+    grep -E "^urn:uuid:${BASE_UUID_REGEX}$"   "$1" > /dev/null || \
+    grep -E "^[[:xdigit:]]{32}$"              "$1" > /dev/null
+}
+
+
 # Check if balena-engine-daemon is responding.
 curl --fail --unix-socket $BALENAD_SOCKET http:/v1.40/_ping > /dev/null 2>&1
 
 # Due to a non-atomic file creation and writing operation in containerd, we
-# sometimes end up with an empty `uuid` file. This causes `ctr version` (and
-# hence the health check) to fail. We therefore remove this file if it is
-# present and empty. See https://github.com/balena-os/balena-engine/issues/322
+# sometimes end up with an empty or corrupted `uuid` file. This causes
+# `ctr version` (and hence the health check) to fail. We therefore remove this
+# file if it is present not valid. See
+# https://github.com/balena-os/balena-engine/issues/322
+
 UUID_FILE="/mnt/data/docker/containerd/daemon/io.containerd.grpc.v1.introspection/uuid"
-if [ -f "$UUID_FILE" -a ! -s "$UUID_FILE" ]; then
-    warn "removing empty $UUID_FILE"
+if [ -f "$UUID_FILE" ] && ! contains_valid_uuid "$UUID_FILE"; then
+    warn "removing invalid $UUID_FILE"
     rm -f "$UUID_FILE"
 fi
 
 # Check if balena-engine-containerd is responding.
 balena-engine-containerd-ctr --address $CONTAINERD_SOCKET version > /dev/null 2>&1
 
-# The uuid file is expected to exist and be non-empty after `ctr version`. If
+# The uuid file is expected to exist and be valid after `ctr version`. If
 # this is not the case, log the event.
-if [ -f "$UUID_FILE" -a ! -s "$UUID_FILE" ]; then
-    warn "$UUID_FILE empty after 'ctr version'"
+if [ -f "$UUID_FILE" ] && ! contains_valid_uuid "$UUID_FILE"; then
+    warn "$UUID_FILE invalid after 'ctr version'"
 fi