Module: check_mk
Branch: master
Commit: 85dd6acf4530af7a5865dd31a14731043bcce2a3
URL: http://git.mathias-kettner.de/git/?p=check_mk.git;a=commit;h=85dd6acf4530af…
Author: Lars Michelsen <lm(a)mathias-kettner.de>
Date: Tue May 22 16:01:38 2018 +0200
6151 FIX Net-SNMP: Add some patch trying to fix SNMPv3 communication in some cases
When using SNMPv3 communication issues with some devices may occur. From the user point of
view timeouts are reported by Check_MK originating from Net-SNMP not receiving a valid response.
This kind of issue is related to a broken negotiation (related to the maximum message size) between
the monitoring and the monitored device.
Change-Id: I310a0f59867130183cc5184bc2195f109c306776
---
.werks/6151 | 14 ++++++++++++++
omd/packages/net-snmp/Makefile | 2 +-
omd/packages/net-snmp/net-snmp-0b32548.tar.gz | Bin 0 -> 6530869 bytes
omd/packages/net-snmp/net-snmp-734f855.tar.gz | Bin 6451953 -> 0 bytes
4 files changed, 15 insertions(+), 1 deletion(-)
diff --git a/.werks/6151 b/.werks/6151
new file mode 100644
index 0000000..100290c
--- /dev/null
+++ b/.werks/6151
@@ -0,0 +1,14 @@
+Title: Net-SNMP: Add some patch trying to fix SNMPv3 communication in some cases
+Level: 1
+Component: packages
+Compatible: compat
+Edition: cre
+Version: 1.6.0i1
+Date: 1526997524
+Class: fix
+
+When using SNMPv3 communication issues with some devices may occur. From the user point of
+view timeouts are reported by Check_MK originating from Net-SNMP not receiving a valid response.
+
+This kind of issue is related to a broken negotiation (related to the maximum message size) between
+the monitoring and the monitored device.
diff --git a/omd/packages/net-snmp/Makefile b/omd/packages/net-snmp/Makefile
index 5ef4ac0..6127b2c 100644
--- a/omd/packages/net-snmp/Makefile
+++ b/omd/packages/net-snmp/Makefile
@@ -1,7 +1,7 @@
include ../../Makefile.omd
NAME = net-snmp
-VERSION = 734f855
+VERSION = 0b32548
DIR = $(NAME)-$(VERSION)
.PHONY: skel
diff --git a/omd/packages/net-snmp/net-snmp-0b32548.tar.gz b/omd/packages/net-snmp/net-snmp-0b32548.tar.gz
new file mode 100644
index 0000000..70a1899
Binary files /dev/null and b/omd/packages/net-snmp/net-snmp-0b32548.tar.gz differ
diff --git a/omd/packages/net-snmp/net-snmp-734f855.tar.gz b/omd/packages/net-snmp/net-snmp-734f855.tar.gz
deleted file mode 100644
index bc48560..0000000
Binary files a/omd/packages/net-snmp/net-snmp-734f855.tar.gz and /dev/null differ
Module: check_mk
Branch: master
Commit: 91a73809de4d34418985263413a794a984543d68
URL: http://git.mathias-kettner.de/git/?p=check_mk.git;a=commit;h=91a73809de4d34…
Author: Lars Michelsen <lm(a)mathias-kettner.de>
Date: Tue May 22 11:57:48 2018 +0200
6149 FIX ps: The linux ps check CPU utilization calculation changed since 1.4.0p27
The calculation of the CPU utilization of processes on linux systems was not taking the number
of CPU cores into account for a long time. This results in CPU utilizations reported with more
than 100%, up to 100%*CPU-cores. For example in case 3 processes use 3 of 6 cores at 100%, this
resulted in a usage of 300%.
The intention of the change in 1.4.0p27 was to make the windows and linux ps checks both take
the number of CPU cores into account and cap the usage at 100%. The problem with this change is
that all possibly existing thresholds that have been configured for the old scheme will not
work anymore with the new scheme.
We have now reverted the changes version 1.4.0p27 introduced to get the old behavior back.
In a future release, we will finally make the switch to the new scheme. But this will be done
in a compatible way where the already configured thresholds will still work.
Change-Id: Idac9be3065a594958125378a6de8e231b8b81467
---
.werks/6149 | 24 ++++++++++++++++++++++++
1 file changed, 24 insertions(+)
diff --git a/.werks/6149 b/.werks/6149
new file mode 100644
index 0000000..7917062
--- /dev/null
+++ b/.werks/6149
@@ -0,0 +1,24 @@
+Title: ps: The linux ps check CPU utilization calculation changed since 1.4.0p27
+Level: 1
+Component: checks
+Class: fix
+Compatible: compat
+Edition: cre
+State: unknown
+Version: 1.6.0i1
+Date: 1526982355
+
+The calculation of the CPU utilization of processes on linux systems was not taking the number
+of CPU cores into account for a long time. This results in CPU utilizations reported with more
+than 100%, up to 100%*CPU-cores. For example in case 3 processes use 3 of 6 cores at 100%, this
+resulted in a usage of 300%.
+
+The intention of the change in 1.4.0p27 was to make the windows and linux ps checks both take
+the number of CPU cores into account and cap the usage at 100%. The problem with this change is
+that all possibly existing thresholds that have been configured for the old scheme will not
+work anymore with the new scheme.
+
+We have now reverted the changes version 1.4.0p27 introduced to get the old behavior back.
+
+In a future release, we will finally make the switch to the new scheme. But this will be done
+in a compatible way where the already configured thresholds will still work.
Module: check_mk
Branch: master
Commit: ee72f0baac0c00df9a63f9e3f63ba6cbd13d9823
URL: http://git.mathias-kettner.de/git/?p=check_mk.git;a=commit;h=ee72f0baac0c00…
Author: Lars Michelsen <lm(a)mathias-kettner.de>
Date: Tue May 22 14:59:03 2018 +0200
6150 FIX docker_container_status: Fix crash when receiving data from multiple hosts
Change-Id: I75e20a53769a7b0f420671d979e711bed68c1ae5
---
.werks/6150 | 11 +++++++++++
checks/docker_container_status | 12 ++++++++++--
2 files changed, 21 insertions(+), 2 deletions(-)
diff --git a/.werks/6150 b/.werks/6150
new file mode 100644
index 0000000..b31a42f
--- /dev/null
+++ b/.werks/6150
@@ -0,0 +1,11 @@
+Title: docker_container_status: Fix crash when receiving data from multiple hosts
+Level: 1
+Component: checks
+Class: fix
+Compatible: compat
+Edition: cre
+State: unknown
+Version: 1.6.0i1
+Date: 1526993923
+
+
diff --git a/checks/docker_container_status b/checks/docker_container_status
index f1cb43d..09e47e4 100644
--- a/checks/docker_container_status
+++ b/checks/docker_container_status
@@ -28,9 +28,17 @@
def parse_docker_container_status(info):
+ if not info:
+ return
+
+ # In case there are multiple lines of output sent by the agent only process the first
+ # line. We assume that this a full JSON object. The rest of the section is skipped.
+ # When a container got piggyback data from multiple hosts (e.g. a cluster) this results
+ # in multiple JSON objects handed over to this check.
+ json_raw_data = info[0]
+
import json
- json_data = "\n".join(" ".join(line) for line in info)
- return json.loads(json_data)
+ return json.loads(json_raw_data)
#.