[reboot] User-friendly reboot cause message for kernel panic (sonic-net#1486)

Signed-off-by: Yong Zhao [email protected]

What I did
If the rebooting of SONiC device was caused by kernel panic, then the CLI command show reboot-cause should show Kernel Panic.

How I did it
Currently if kernel was panicked, then the device would be rebooted. The reboot script wrote a message into reboot-cause.txt. I just updated the content of this message.

How to verify it
I verified this change on the virtual switch in the following steps:

Trigger kernel panic: echo c > /proc/sysrq-trigger
After device was rebooted, run the CLI show reboot-cause:
admin@vlab-01:~$ show reboot-cause
Kernel Panic [Time: Tue 09 Mar 2021 03:03:56 AM UTC]
Previous command output (if the output of a command-line utility has changed)
admin@vlab-01:~$ show reboot-cause
User issued 'kdump' command [User: kdump, Time: Mon 08 Mar 2021 01:47:43 AM UTC]

New command output (if the output of a command-line utility has changed)
admin@vlab-01:~$ show reboot-cause
Kernel Panic [Time: Tue 09 Mar 2021 03:03:56 AM UTC]

Author: yozhao101

scripts/reboot

@@ -9,7 +9,7 @@ REBOOT_TIME=$(date)
 VMCORE_FILE=/proc/vmcore
 if [ -e $VMCORE_FILE -a -s $VMCORE_FILE ]; then
         echo "We have a /proc/vmcore, then we just kdump'ed"
-        echo "User issued 'kdump' command [User: kdump, Time: ${REBOOT_TIME}]" > ${REBOOT_CAUSE_FILE}
+        echo "Kernel Panic [Time: ${REBOOT_TIME}]" > ${REBOOT_CAUSE_FILE}
         sync
         PLATFORM=$(grep -oP 'sonic_platform=\K\S+' /proc/cmdline)
         if [ ! -z "${PLATFORM}" -a -x ${DEVPATH}/${PLATFORM}/${PLAT_REBOOT} ]; then

show/reboot_cause.py

@@ -8,15 +8,21 @@
 import utilities_common.cli as clicommon
 
 
-PREVIOUS_REBOOT_CAUSE_FILE = "/host/reboot-cause/previous-reboot-cause.json"
-USER_ISSUED_REBOOT_CAUSE_REGEX ="User issued \'{}\' command [User: {}, Time: {}]"
+PREVIOUS_REBOOT_CAUSE_FILE_PATH = "/host/reboot-cause/previous-reboot-cause.json"
+
 
 def read_reboot_cause_file():
-    result = ""
-    if os.path.exists(PREVIOUS_REBOOT_CAUSE_FILE):
-        with open(PREVIOUS_REBOOT_CAUSE_FILE) as f:
-            result = json.load(f)
-    return result
+    reboot_cause_dict = {}
+
+    if os.path.exists(PREVIOUS_REBOOT_CAUSE_FILE_PATH):
+        with open(PREVIOUS_REBOOT_CAUSE_FILE_PATH) as prev_reboot_cause_file:
+            try:
+                reboot_cause_dict = json.load(prev_reboot_cause_file)
+            except json.JSONDecodeError as err:
+                click.echo("Failed to load JSON file '{}'!".format(PREVIOUS_REBOOT_CAUSE_FILE_PATH), err=True)
+
+    return reboot_cause_dict
+
 
 #
 # 'reboot-cause' group ("show reboot-cause")
@@ -26,15 +32,35 @@ def read_reboot_cause_file():
 def reboot_cause(ctx):
     """Show cause of most recent reboot"""
     if ctx.invoked_subcommand is None:
-        reboot_cause = ""
+        reboot_cause_str = ""
+
         # Read the previous reboot cause
-        data = read_reboot_cause_file()
-        if data['user'] == "N/A":
-            reboot_cause = "{}".format(data['cause'])
+        reboot_cause_dict = read_reboot_cause_file()
+
+        reboot_cause = reboot_cause_dict.get("cause", "Unknown")
+        reboot_user = reboot_cause_dict.get("user", "N/A")
+        reboot_time = reboot_cause_dict.get("time", "N/A")
+
+        if reboot_user != "N/A":
+            reboot_cause_str = "User issued '{}' command".format(reboot_cause)
         else:
-            reboot_cause = USER_ISSUED_REBOOT_CAUSE_REGEX.format(data['cause'], data['user'], data['time'])
+            reboot_cause_str = reboot_cause
+
+        if reboot_user != "N/A" or reboot_time != "N/A":
+            reboot_cause_str += " ["
+
+            if reboot_user != "N/A":
+                reboot_cause_str += "User: {}".format(reboot_user)
+                if reboot_time != "N/A":
+                    reboot_cause_str += ", "
+
+            if reboot_time != "N/A":
+                reboot_cause_str += "Time: {}".format(reboot_time)
+
+            reboot_cause_str += "]"
+
+        click.echo(reboot_cause_str)
 
-        click.echo(reboot_cause)
 
 # 'history' subcommand ("show reboot-cause history")
 @reboot_cause.command()
@@ -54,7 +80,7 @@ def history():
         for tk in table_keys:
             entry = db.get_all(db.STATE_DB, tk)
             r = []
-            r.append(tk.replace(prefix,""))
+            r.append(tk.replace(prefix, ""))
             r.append(entry['cause'] if 'cause' in entry else "")
             r.append(entry['time'] if 'time' in entry else "")
             r.append(entry['user'] if 'user' in entry else "")

tests/reboot_cause_test.py

@@ -31,7 +31,7 @@ def setup_class(cls):
 
     # Test 'show reboot-cause' without previous-reboot-cause.json 
     def test_reboot_cause_no_history_file(self):
-        expected_output = ""
+        expected_output = "Unknown\n"
         runner = CliRunner()
         result = runner.invoke(show.cli.commands["reboot-cause"], [])
         assert result.output == expected_output