Module: check_mk Branch: master Commit: be5fbb1e34cbc625553ebf9c717cd4f7fe4e3fde URL: http://git.mathias-kettner.de/git/?p=check_mk.git;a=commit;h=be5fbb1e34cbc6… Author: Moritz Kiemer <mo(a)mathias-kettner.de> Date: Fri Aug 31 13:49:33 2018 +0200 6218 FIX mgmt_ipmi_sensors: Wrong battery / power supply WARN state Due to a bug in pyghmi (https://bugs.launchpad.net/pyghmi/+bug/1790120) a WARN state was reported for sensors of type "battery" and "power supply". A workaroud has been added to mitigate this problem. FEED-943 Change-Id: I9af116870b977706305ab9ae6253b8c0d61d7893 --- .werks/6218 | 12 ++++++++++++ checks/ipmi_sensors.include | 8 ++++---- cmk_base/data_sources/ipmi.py | 34 ++++++++++++++++++++++++++++++++++ 3 files changed, 50 insertions(+), 4 deletions(-) diff --git a/.werks/6218 b/.werks/6218 new file mode 100644 index 0000000..81c0695 --- /dev/null +++ b/.werks/6218 @@ -0,0 +1,12 @@ +Title: mgmt_ipmi_sensors: Wrong battery / power supply WARN state +Level: 1 +Component: checks +Compatible: compat +Edition: cre +Version: 1.6.0i1 +Date: 1535715823 +Class: fix + +Due to a bug in pyghmi (https://bugs.launchpad.net/pyghmi/+bug/1790120) +a WARN state was reported for sensors of type "battery" and "power supply". +A workaroud has been added to mitigate this problem. diff --git a/checks/ipmi_sensors.include b/checks/ipmi_sensors.include index 7b36483..adaf1d2 100644 --- a/checks/ipmi_sensors.include +++ b/checks/ipmi_sensors.include @@ -198,10 +198,10 @@ def freeipmi_status_txt_mapping(status_txt): if state is not None: return state - if status_txt in [ - "Entity Present", "battery presence detected", - "Drive Presence", "transition to Running", "Device Enabled", - "System full operational, working", "System Restart", + if status_txt.lower() in [ + "entity present", "battery presence detected", + "drive presence", "transition to running", "device enabled", + "system full operational, working", "system restart", "present", ] or \ status_txt.startswith("Fully Redundant") or \ status_txt.endswith("is connected") or \ diff --git a/cmk_base/data_sources/ipmi.py b/cmk_base/data_sources/ipmi.py index 4b09bcc..1f23c93 100644 --- a/cmk_base/data_sources/ipmi.py +++ b/cmk_base/data_sources/ipmi.py @@ -37,6 +37,38 @@ from cmk_base.exceptions import MKAgentError from .abstract import CheckMKAgentDataSource, ManagementBoardDataSource +def _handle_false_positive_warnings(reading): + """This is a workaround for a pyghmi bug + (bug report: https://bugs.launchpad.net/pyghmi/+bug/1790120) + + For some sensors undefined states are looked up, which results in readings of the form + {'states': ['Present', + 'Unknown state 8 for reading type 111/sensor type 8', + 'Unknown state 9 for reading type 111/sensor type 8', + 'Unknown state 10 for reading type 111/sensor type 8', + 'Unknown state 11 for reading type 111/sensor type 8', + 'Unknown state 12 for reading type 111/sensor type 8', ...], + 'health': 1, 'name': 'PS Status', 'imprecision': None, 'units': '', + 'state_ids': [552704, 552712, 552713, 552714, 552715, 552716, 552717, 552718], + 'type': 'Power Supply', 'value': None, 'unavailable': 0} + + The health warning is set, but only due to the lookup errors. We remove the lookup + errors, and see whether the remaining states are meaningful. + """ + states = [s for s in reading.states if not s.startswith("Unknown state ")] + + if not states: + return "no state reported" + + if any("non-critical" in s for s in states): + return "WARNING" + + # just keep all the available info. It should be dealt with in + # ipmi_sensors.include (freeipmi_status_txt_mapping), + # where it will default to 2(CRIT) + return ', '.join(states) + + class IPMIManagementBoardDataSource(ManagementBoardDataSource, CheckMKAgentDataSource): def id(self): @@ -126,6 +158,8 @@ class IPMIManagementBoardDataSource(ManagementBoardDataSource, CheckMKAgentDataS health_txt = "CRITICAL" elif reading.health >= ipmi_const.Health.Warning: health_txt = "WARNING" + # workaround for pyghmi bug: https://bugs.launchpad.net/pyghmi/+bug/1790120 + health_txt = _handle_false_positive_warnings(reading) elif reading.health == ipmi_const.Health.Ok: health_txt = "OK"