[PATCH] HP SmartArray plugin: Improve hpssacli output parsing algorithm.

[Gris Ge]

* The output of hpssacli contains incorrect indented section which
cause 'duplicate line' PLUGIN_BUG error. To fix it, we only
parse required section.

* Add more comments to explain the workflow of parsing hpssacli output.

* Tested on SmartArray with and without disk enclosure.

Signed-off-by: Gris Ge <***@redhat.com>
---
plugin/hpsa/hpsa.py | 48 +++++++++++++++++++++++++++++++++---------------
1 file changed, 33 insertions(+), 15 deletions(-)

diff --git a/plugin/hpsa/hpsa.py b/plugin/hpsa/hpsa.py
index 42e0bbb..6b273c9 100644
--- a/plugin/hpsa/hpsa.py
+++ b/plugin/hpsa/hpsa.py
@@ -74,38 +74,38 @@ def _sys_status_of(hp_ctrl_status):
def _parse_hpssacli_output(output):
"""
Got a output string of hpssacli to dictionary(nested).
- Skipped these line:
- 1. Starts with 'Note:'
- This is just a message right after controller. We don't neet it
- yet.
- 2. The 'Physical Drives' line.
- It should indented after 'Internal Drive Cage' like.
- If not ignored, we might got duplication line error.
- After ignored, it's phsycial disks will directly stored as
- key of 'Internal Drive Cage' dictionary.
+ Workflow:
+ 0. Check out top and the second indention level space count.
+ 1. Check current line and next line to determine whether current line is
+ a start of new setction.
+ 2. Skip all unrequired sections and their data and sub-sections.
+ 3. If current line is the start of new section, create and empty dict
+ where following subsection or data could be stored.
"""
+ required_sections = ['Array:', 'unassigned']
+
output_lines = [
l for l in output.split("\n")
- if l and not l.startswith('Note:') and
- not l.strip() == 'Physical Drives']
+ if l and not l.startswith('Note:')]

data = {}

# Detemine indention level
- top_indention_level = sorted(
+ (top_indention_level, second_indention_level) = sorted(
set(
len(line) - len(line.lstrip())
- for line in output_lines))[0]
+ for line in output_lines))[0:2]

indent_2_data = {
top_indention_level: data
}

+ flag_required_section = False
+
for line_num in range(len(output_lines)):
cur_line = output_lines[line_num]
- if cur_line.strip() == 'None attached':
- continue
if line_num + 1 == len(output_lines):
+ # The current line is the last line.
nxt_line = ''
else:
nxt_line = output_lines[line_num + 1]
@@ -113,10 +113,27 @@ def _parse_hpssacli_output(output):
cur_indent_count = len(cur_line) - len(cur_line.lstrip())
nxt_indent_count = len(nxt_line) - len(nxt_line.lstrip())

+ if cur_indent_count == top_indention_level:
+ flag_required_section = True
+
+ if cur_indent_count == second_indention_level:
+ flag_required_section = False
+
+ if nxt_indent_count == cur_indent_count:
+ flag_required_section = True
+ else:
+ for required_section in required_sections:
+ if cur_line.lstrip().startswith(required_section):
+ flag_required_section = True
+
+ if flag_required_section is False:
+ continue
+
cur_line_splitted = cur_line.split(": ")
cur_data_pointer = indent_2_data[cur_indent_count]

if nxt_indent_count > cur_indent_count:
+ # Current line is new section title
nxt_line_splitted = nxt_line.split(": ")
new_data = {}

@@ -134,6 +151,7 @@ def _parse_hpssacli_output(output):
else:
cur_data_pointer[cur_line_splitted[0].lstrip()] = \
": ".join(cur_line_splitted[1:]).strip()
+
return data

[Gris Ge]

* The output of hpssacli contains incorrect indented section which
cause 'duplicate line' PLUGIN_BUG error. To fix it, we only
parse required section.

* Add more comments to explain the workflow of parsing hpssacli output.

* Tested on SmartArray with and without disk enclosure.

Changes in V2:

* Fix spell errors.

Signed-off-by: Gris Ge <***@redhat.com>
---
plugin/hpsa/hpsa.py | 50 ++++++++++++++++++++++++++++++++++----------------
1 file changed, 34 insertions(+), 16 deletions(-)

diff --git a/plugin/hpsa/hpsa.py b/plugin/hpsa/hpsa.py
index 42e0bbb..4a3d373 100644
--- a/plugin/hpsa/hpsa.py
+++ b/plugin/hpsa/hpsa.py
@@ -74,38 +74,38 @@ def _sys_status_of(hp_ctrl_status):
def _parse_hpssacli_output(output):
"""
Got a output string of hpssacli to dictionary(nested).
- Skipped these line:
- 1. Starts with 'Note:'
- This is just a message right after controller. We don't neet it
- yet.
- 2. The 'Physical Drives' line.
- It should indented after 'Internal Drive Cage' like.
- If not ignored, we might got duplication line error.
- After ignored, it's phsycial disks will directly stored as
- key of 'Internal Drive Cage' dictionary.
+ Workflow:
+ 0. Check out top and the second indention level's space count.
+ 1. Check current line and next line to determine whether current line is
+ a start of new section.
+ 2. Skip all un-required sections and their data and sub-sections.
+ 3. If current line is the start of new section, create an empty dictionary
+ where following subsections or data could be stored in.
"""
+ required_sections = ['Array:', 'unassigned']
+
output_lines = [
l for l in output.split("\n")
- if l and not l.startswith('Note:') and
- not l.strip() == 'Physical Drives']
+ if l and not l.startswith('Note:')]

data = {}

- # Detemine indention level
- top_indention_level = sorted(
+ # Determine indention level
+ (top_indention_level, second_indention_level) = sorted(
set(
len(line) - len(line.lstrip())
- for line in output_lines))[0]
+ for line in output_lines))[0:2]

indent_2_data = {
top_indention_level: data
}

+ flag_required_section = False
+
for line_num in range(len(output_lines)):
cur_line = output_lines[line_num]
- if cur_line.strip() == 'None attached':
- continue
if line_num + 1 == len(output_lines):
+ # The current line is the last line.
nxt_line = ''
else:
nxt_line = output_lines[line_num + 1]
@@ -113,10 +113,27 @@ def _parse_hpssacli_output(output):
cur_indent_count = len(cur_line) - len(cur_line.lstrip())
nxt_indent_count = len(nxt_line) - len(nxt_line.lstrip())

+ if cur_indent_count == top_indention_level:
+ flag_required_section = True
+
+ if cur_indent_count == second_indention_level:
+ flag_required_section = False
+
+ if nxt_indent_count == cur_indent_count:
+ flag_required_section = True
+ else:
+ for required_section in required_sections:
+ if cur_line.lstrip().startswith(required_section):
+ flag_required_section = True
+
+ if flag_required_section is False:
+ continue
+
cur_line_splitted = cur_line.split(": ")
cur_data_pointer = indent_2_data[cur_indent_count]

if nxt_indent_count > cur_indent_count:
+ # Current line is new section title
nxt_line_splitted = nxt_line.split(": ")
new_data = {}

@@ -134,6 +151,7 @@ def _parse_hpssacli_output(output):
else:
cur_data_pointer[cur_line_splitted[0].lstrip()] = \
": ".join(cur_line_splitted[1:]).strip()
+
return data

[Gris Ge]

* Replace variable 'cur_line_splitted' with 'cur_line_split'.

Signed-off-by: Gris Ge <***@redhat.com>
---
plugin/hpsa/hpsa.py | 10 +++++-----
1 file changed, 5 insertions(+), 5 deletions(-)

diff --git a/plugin/hpsa/hpsa.py b/plugin/hpsa/hpsa.py
index 4a3d373..9c49cbe 100644
--- a/plugin/hpsa/hpsa.py
+++ b/plugin/hpsa/hpsa.py
@@ -129,12 +129,12 @@ def _parse_hpssacli_output(output):
if flag_required_section is False:
continue

- cur_line_splitted = cur_line.split(": ")
+ cur_line_split = cur_line.split(": ")
cur_data_pointer = indent_2_data[cur_indent_count]

if nxt_indent_count > cur_indent_count:
# Current line is new section title
- nxt_line_splitted = nxt_line.split(": ")
+ nxt_line_split = nxt_line.split(": ")
new_data = {}

if cur_line.lstrip() not in cur_data_pointer:
@@ -146,11 +146,11 @@ def _parse_hpssacli_output(output):
"_parse_hpssacli_output(): Found duplicate line %s" %
cur_line)
else:
- if len(cur_line_splitted) == 1:
+ if len(cur_line_split) == 1:
cur_data_pointer[cur_line.lstrip()] = None
else:
- cur_data_pointer[cur_line_splitted[0].lstrip()] = \
- ": ".join(cur_line_splitted[1:]).strip()
+ cur_data_pointer[cur_line_split[0].lstrip()] = \
+ ": ".join(cur_line_split[1:]).strip()

return data