UnicodeDecodeError: 'utf-8' codec can't decode bytes in position 146-147: invalid continuation byte about hwraid

[eLvErDe]

Good 'ol iso8859 :D

I'll send you a patch to try tomorrow

[eLvErDe]

It seems your megacli outputs some characters in a non-utf8 charset, that's very weird.
Can you send the output to a file and run "file" on it to see what charset it detects ?

[Delido]

/usr/sbin/megacli -LDInfo -l0 -a0 -NoLog > /tmp/mega
❯ file -0 /tmp/mega
/tmp/mega: ISO-8859 text, with CR, LF line terminators

on another working Server:

file -0 /tmp/mega
/tmp/mega: ASCII text, with CR, LF line terminators

[Delido]

why is the output iso8859? all locale Settings between the Servers are the same and UTF-8 are default. Can / must i change something?

[krietvel]

We are running a bunch of old PERC 5/i adapters, for which MegaCLI 8.07.14 gives non-UTF8 and non-ASCII output (classified as "data" by the file command). I ended up using the patch below to get megaclisas-status to work with Python 3.9 on bullseye. It replaces popen with subprocess.Popen to be able to ignore the conversion errors ...

diff a/wrapper-scripts/megaclisas-status b/wrapper-scripts/megaclisas-status
--- a/wrapper-scripts/megaclisas-status
+++ b/wrapper-scripts/megaclisas-status
@@ -11,6 +11,7 @@ import sys
 import pdb
 import inspect
 import argparse
+from subprocess import Popen, PIPE, STDOUT
 
 if sys.platform == "win32":
     import ctypes
@@ -144,8 +145,10 @@ def getOutput(cmd):
         lines = Outputs[cmd]
     else:
         dbgprint("Not a Cached value: " + str(cmd))
-        output = os.popen(cmd)
-        for line in output:
+        process = Popen(cmd, shell=True, stdout=PIPE, stderr=STDOUT, text=True, errors='ignore')
+        process.wait()
+        output = process.stdout.read()
+        for line in output.split("\n"):
             if not re.match(r"^$", line.strip()):
                 lines.append(line.strip())
         Outputs[cmd] = lines