From 405a2b2a55866b74f196ad81eaafedeee4615fa8 Mon Sep 17 00:00:00 2001
From: Tim Burke <tim.burke@gmail.com>
Date: Sun, 13 Oct 2019 21:55:54 -0700
Subject: [PATCH] py3: Fix swift-drive-audit

Walking through the kernel logs backwards requires that we open them
in binary mode. Add a new option to allow users to specify which
encoding should be used to interpret those logs; default to the same
encoding that open() uses for its default.

Change-Id: Iae332bb58388b5521445e75beba6ee2e9f06bfa6
Closes-Bug: #1847955
---
 bin/swift-drive-audit       | 17 +++++++++++++----
 etc/drive-audit.conf-sample |  4 ++++
 2 files changed, 17 insertions(+), 4 deletions(-)

diff --git a/bin/swift-drive-audit b/bin/swift-drive-audit
index bda4cd2af7..c9c456d3af 100755
--- a/bin/swift-drive-audit
+++ b/bin/swift-drive-audit
@@ -16,12 +16,14 @@
 
 import datetime
 import glob
+import locale
 import os
 import re
 import subprocess
 import sys
 
 
+import six
 from six.moves.configparser import ConfigParser
 
 from swift.common.utils import backward, get_logger, dump_recon_cache, \
@@ -57,7 +59,8 @@ def get_devices(device_dir, logger):
     return devices
 
 
-def get_errors(error_re, log_file_pattern, minutes, logger):
+def get_errors(error_re, log_file_pattern, minutes, logger,
+               log_file_encoding):
     # Assuming log rotation is being used, we need to examine
     # recently rotated files in case the rotation occurred
     # just before the script is being run - the data we are
@@ -87,12 +90,14 @@ def get_errors(error_re, log_file_pattern, minutes, logger):
     reached_old_logs = False
     for path in log_files:
         try:
-            f = open(path)
+            f = open(path, 'rb')
         except IOError:
             logger.error("Error: Unable to open " + path)
             print("Unable to open " + path)
             sys.exit(1)
         for line in backward(f):
+            if not six.PY2:
+                line = line.decode(log_file_encoding, 'surrogateescape')
             if '[    0.000000]' in line \
                 or 'KERNEL supported cpus:' in line \
                     or 'BIOS-provided physical RAM map:' in line:
@@ -105,7 +110,7 @@ def get_errors(error_re, log_file_pattern, minutes, logger):
             if log_time_entry[0] == 'Dec' and prev_entry_month == 'Jan':
                 year -= 1
             prev_entry_month = log_time_entry[0]
-            log_time_string = '%s %s' % (year, ' '.join(log_time_entry))
+            log_time_string = '%d %s' % (year, ' '.join(log_time_entry))
             try:
                 log_time = datetime.datetime.strptime(
                     log_time_string, '%Y %b %d %H:%M:%S')
@@ -154,6 +159,9 @@ if __name__ == '__main__':
     recon_cache_path = conf.get('recon_cache_path', "/var/cache/swift")
     log_file_pattern = conf.get('log_file_pattern',
                                 '/var/log/kern.*[!.][!g][!z]')
+    log_file_encoding = conf.get('log_file_encoding', 'auto')
+    if log_file_encoding == 'auto':
+        log_file_encoding = locale.getpreferredencoding()
     log_to_console = config_true_value(conf.get('log_to_console', False))
     error_re = []
     for conf_key in conf:
@@ -181,7 +189,8 @@ if __name__ == '__main__':
     total_errors = 0
     for device in devices:
         recon_errors[device['mount_point']] = 0
-    errors = get_errors(error_re, log_file_pattern, minutes, logger)
+    errors = get_errors(error_re, log_file_pattern, minutes, logger,
+                        log_file_encoding)
     logger.debug("Errors found: %s" % str(errors))
     unmounts = 0
     for kernel_device, count in errors.items():
diff --git a/etc/drive-audit.conf-sample b/etc/drive-audit.conf-sample
index 7f16517cb3..69de6defbf 100644
--- a/etc/drive-audit.conf-sample
+++ b/etc/drive-audit.conf-sample
@@ -26,6 +26,10 @@
 # pattern to check against device errors.
 # log_file_pattern = /var/log/kern.*[!.][!g][!z]
 #
+# On Python 3, the encoding to use when reading the log file. Defaults
+# to the result of locale.getpreferredencoding(), like Python's open().
+# log_file_encoding = auto
+#
 # Regular expression patterns to be used to locate
 # device blocks with errors in the log file. Currently
 # the default ones are as follows: