[Check_mk (english)] windows check_mk_agent 1.2.8p12 with hpacucli.exe

[WoutR (Wouter Bleeker)]

So how is your agent started in your situation? You push the results, so it doesn’t run as a service?

Can you get a regular powershell script to write-out any output? When you add something like “add-content full-path-to-logfile.log test_txt” in the first line, is your script running at all?

···

Op 15 okt. 2016 om 09:05 heeft Pavel Vaysman goam03@gmail.com het volgende geschreven:

I have about 300 server behind firewalls and nat and it’s impossible to query check_mk agent and snmp from central check_mk server.

So i run check_mk agent locally on server and push it’s output to check_mk server.

On Sat, Oct 15, 2016 at 12:05 AM, mlist@libero.it mlist@libero.it wrote:

Pavel,

why are you using a powershell script? in my test environment I added some proliant dl servers running linux + hp agents and it worked out of the box. I did not try windows servers but it should be the same; I just selected agent+snmp

----Messaggio originale----

Da: “Pavel Vaysman” goam03@gmail.com

Data: 14/10/2016 17.27

A: checkmk-en@lists.mathias-kettner.de

Ogg: [Check_mk (english)] windows check_mk_agent 1.2.8p12 with hpacucli.exe

Hi guys,

i try to monitor hp raid with mrpe via check_mk_agent 1.2.8p12 with hpacucli.exe.
I have powershell script that run hpacucli.exe and parsing it’s output.
With check_mk_agent 1.2.6p16 it’s working perfect.
With new check_mk agent 1.2.8pX it’s return “ERROR: Cannot open console”
When i’m run check_mk agent 1.2.8pX with debug argument - it’s working.
Right now , workaround - write hpacucli.exe output to file via windows task scheduler and parse this txt file.

Is it bug or may be i did mistake somewhere?

Source Code

1. $array_config_util = “c:\hp\Bin\hpacucli.exe”
2. try {
3. $exec = & $array_config_util ‘ctrl all show config’
4. #write-host !!!
5. #Write-Output $exec
6. }
7. catch {
8. Write-Host “Problem checking array status)”
9. exit 3
10. }
11. # filter results for lines that talk about drives (physicaldrive, logicaldrive)
12. # and do not end with “OK”:
13. $not_OK = $exec | Where-Object { $_ -like “drive” } | Where-Object { $_ -notlike “OK)” }
14. #write-output “<<<>>>”
15. #write-output $not_OK
16. if ($not_OK.length -lt 1) {
17. Write-Host “Array status appears OK”
18. exit 0
19. }
20. Write-Host “Array status not OK; please check RAID software”
21. exit 2

[Goam03 (Pavel Vaysman)]

i push result with external python program.

https://github.com/qmexnetworks/cmk-pusher-agent

but it’s not powershell problem…

when i run this script manually from powershell it’s working, from debug mode in cmd “check_mk_agent debug” its working , but “check_mk_agent test” - it failed with error…

this mrpe check is working with agent 1.2.6p12…

···

On Oct 16, 2016 15:06, “Wouter” wbleeker@gmail.com wrote:

So how is your agent started in your situation? You push the results, so it doesn’t run as a service?

Can you get a regular powershell script to write-out any output? When you add something like “add-content full-path-to-logfile.log test_txt” in the first line, is your script running at all?

Verstuurd vanaf mijn iPhone

Op 15 okt. 2016 om 09:05 heeft Pavel Vaysman goam03@gmail.com het volgende geschreven:

I have about 300 server behind firewalls and nat and it’s impossible to query check_mk agent and snmp from central check_mk server.

So i run check_mk agent locally on server and push it’s output to check_mk server.

On Sat, Oct 15, 2016 at 12:05 AM, mlist@libero.it mlist@libero.it wrote:

Pavel,

why are you using a powershell script? in my test environment I added some proliant dl servers running linux + hp agents and it worked out of the box. I did not try windows servers but it should be the same; I just selected agent+snmp

----Messaggio originale----

Da: “Pavel Vaysman” goam03@gmail.com

Data: 14/10/2016 17.27

A: checkmk-en@lists.mathias-kettner.de

Ogg: [Check_mk (english)] windows check_mk_agent 1.2.8p12 with hpacucli.exe

Hi guys,

i try to monitor hp raid with mrpe via check_mk_agent 1.2.8p12 with hpacucli.exe.
I have powershell script that run hpacucli.exe and parsing it’s output.
With check_mk_agent 1.2.6p16 it’s working perfect.
With new check_mk agent 1.2.8pX it’s return “ERROR: Cannot open console”
When i’m run check_mk agent 1.2.8pX with debug argument - it’s working.
Right now , workaround - write hpacucli.exe output to file via windows task scheduler and parse this txt file.

Is it bug or may be i did mistake somewhere?

Source Code

1. $array_config_util = “c:\hp\Bin\hpacucli.exe”
2. try {
3. $exec = & $array_config_util ‘ctrl all show config’
4. #write-host !!!
5. #Write-Output $exec
6. }
7. catch {
8. Write-Host “Problem checking array status)”
9. exit 3
10. }
11. # filter results for lines that talk about drives (physicaldrive, logicaldrive)
12. # and do not end with “OK”:
13. $not_OK = $exec | Where-Object { $_ -like “drive” } | Where-Object { $_ -notlike “OK)” }
14. #write-output “<<<>>>”
15. #write-output $not_OK
16. if ($not_OK.length -lt 1) {
17. Write-Host “Array status appears OK”
18. exit 0
19. }
20. Write-Host “Array status not OK; please check RAID software”
21. exit 2

[WoutR (Wouter Bleeker)]

It is not related to your cmk-pusher script.

···

Can you share the config of check_mk? The regular ini file in which you added the mrpe command?

Also: can you follow up on my question to add the log write action as the first line in your powershell script? It will (probably) show that the powershell script cannot be executed at all. So in that case it is not related to the commands in your scripts.

Verstuurd vanaf mijn iPhone

Op 16 okt. 2016 om 23:21 heeft Pavel Vaysman goam03@gmail.com het volgende geschreven:

i push result with external python program.

https://github.com/qmexnetworks/cmk-pusher-agent

but it’s not powershell problem…

when i run this script manually from powershell it’s working, from debug mode in cmd “check_mk_agent debug” its working , but “check_mk_agent test” - it failed with error…

this mrpe check is working with agent 1.2.6p12…

On Oct 16, 2016 15:06, “Wouter” wbleeker@gmail.com wrote:

So how is your agent started in your situation? You push the results, so it doesn’t run as a service?

Can you get a regular powershell script to write-out any output? When you add something like “add-content full-path-to-logfile.log test_txt” in the first line, is your script running at all?

Verstuurd vanaf mijn iPhone

Op 15 okt. 2016 om 09:05 heeft Pavel Vaysman goam03@gmail.com het volgende geschreven:

I have about 300 server behind firewalls and nat and it’s impossible to query check_mk agent and snmp from central check_mk server.

So i run check_mk agent locally on server and push it’s output to check_mk server.

On Sat, Oct 15, 2016 at 12:05 AM, mlist@libero.it mlist@libero.it wrote:

Pavel,

why are you using a powershell script? in my test environment I added some proliant dl servers running linux + hp agents and it worked out of the box. I did not try windows servers but it should be the same; I just selected agent+snmp

----Messaggio originale----

Da: “Pavel Vaysman” goam03@gmail.com

Data: 14/10/2016 17.27

A: checkmk-en@lists.mathias-kettner.de

Ogg: [Check_mk (english)] windows check_mk_agent 1.2.8p12 with hpacucli.exe

Hi guys,

i try to monitor hp raid with mrpe via check_mk_agent 1.2.8p12 with hpacucli.exe.
I have powershell script that run hpacucli.exe and parsing it’s output.
With check_mk_agent 1.2.6p16 it’s working perfect.
With new check_mk agent 1.2.8pX it’s return “ERROR: Cannot open console”
When i’m run check_mk agent 1.2.8pX with debug argument - it’s working.
Right now , workaround - write hpacucli.exe output to file via windows task scheduler and parse this txt file.

Is it bug or may be i did mistake somewhere?

Source Code

1. $array_config_util = “c:\hp\Bin\hpacucli.exe”
2. try {
3. $exec = & $array_config_util ‘ctrl all show config’
4. #write-host !!!
5. #Write-Output $exec
6. }
7. catch {
8. Write-Host “Problem checking array status)”
9. exit 3
10. }
11. # filter results for lines that talk about drives (physicaldrive, logicaldrive)
12. # and do not end with “OK”:
13. $not_OK = $exec | Where-Object { $_ -like “drive” } | Where-Object { $_ -notlike “OK)” }
14. #write-output “<<<>>>”
15. #write-output $not_OK
16. if ($not_OK.length -lt 1) {
17. Write-Host “Array status appears OK”
18. exit 0
19. }
20. Write-Host “Array status not OK; please check RAID software”
21. exit 2

[Goam03 (Pavel Vaysman)]

Hi,
I added this line to to the script. it’s create “log file” without problem.

this is check_mk.ini file configuration:

Example configuration for Windows agent

[global]

Restrict access to certain IP addresses

If ipv6 is enabled, all listed ipv4 adresses are also accepted as

source adresses in their ipv6-mapped form. I.e. if

192.168.56.0/24 is listed, connections from 0:0:0:0:0:ffff:c0a8:3800/120

are also possible

only_from = 127.0.0.1 192.168.56.0/24 ::1

Change port where the agent is listening ( default 6556 )

port = 6556

Disable ipv6 support. By default, ipv4 is always supported

and ipv6 is enabled additionally if supported by the system.

ipv6 = no

Do only execute programs and scripts with

the following suffixes (in local and plugins).

Per default all extensions except txt and dir

are being executed.

execute = exe bat vbs

Restrict the following configuration variables

in [global] to the following hosts

host = winsrv* zab???ad

Just output certain sections (e.g. for upper hosts)

sections = check_mk winperf

output sections in realtime updates. Please note that not all

sections can be used in realtime updates, the following is the

complete list

realtime_sections = mem df winperf_processor

specifies how long (in seconds) realtime updates are sent to

the last monitoring system that requested an update.

this should be longer than the request frequency (usually

one minute).

Please note that any new request cancels previous realtime

update schedules, so no more than one update is sent per second,

no matter if this timeout is “too high” or how many monitoring

systems are querying the agent.

realtime_timeout = 90

passphrase for encrypted communication. Currently this is only

used in realtime updates

passphrase = secret

Write a logfile for tackling down crashes of the agent

crash_debug = yes

by default the agent flushes the socket for each agent section.

this is mostly for debugging purposes in case the agent is slow

to respond.

Disabling this may (very) slightly reduce network traffic.

section_flush = no

[winperf]

Select counters to extract. The following counters

are needed by checks shipped with check_mk.

counters = 10332:msx_queues

counters = 638:tcp_conn

[logfiles]

# Define textfiles to be monitored, separated by |

textfile = C:\tmp logfiles\message_*.log|D:\log\sample.txt

# Set patterns for defined textfiles

ok = Successfully logged in*

crit = Error in*

warn = Unable to contact*

ignore = Backup * saved

# Define additional textfiles with different patterns

textfile = C:\tmp\memo.udf

# Set patterns for defined textfile

warn = overdue

ok = mail sent

[logwatch]

Testing: output all messages from the eventlogs

sendall = yes

From application log send only critical messages

logfile application = crit

From system log send only warning/critical messages,

but suppress any context messages

logfile system = nocontext warn

Switch all other logfiles off. Default is warn:

send messages of type warn or crit

logfile * = off

[mrpe]

Run classical monitoring plugins. The word before the command

line is the service description for the monitoring. Use backslashes

in Windows-paths.

check = IP_Configuration mrpe\check_ipconfig 1.2.3.4

check = Whatever c:\myplugins\check_whatever -w 10 -c 20

check = HP_Smart_Array c:\Windows\System32\WindowsPowerShell\v1.0\powershell.exe -command “.‘c:\medasys-monitor\mrpe\hp-raid.ps1’; exit($lastexitcode)”

[fileinfo]

path = C:\Programs\Foo*.log

path = M:\Bar Test*.*

[local]

define timeouts for local scripts matching

specific patterns - first match wins

timeout *.vbs = 20

timeout *.bat = 10

timeout * = 30

[plugins]

example: the windows_updates.vbs

plugin is executed asynchronous

and is only updated every 3600 seconds

it may fail (timeout / error) up to 3 times before the last known data is discarded

execution windows_updates.vbs = async

timeout windows_updates.vbs = 120

cache_age windows_updates.vbs = 3600

retry_count windows_udpates.vbs = 3

define timeouts for plugin scripts matching

specific patterns - first match wins

timeout ps_perf.ps1 = 20

timeout *.ps1 = 10

timeout * = 30

When using the Check_MK Inventory plugin, it is a good idea to make the

plugin being executed asynchronous to prevent it from hooking up the

whole agent processing. Additionally it should have a execution timeout.

execution mk_inventory.ps1 = async

timeout mk_inventory.ps1 = 240

[ps]

Experimental: Set to yes to use wmi for retrieving process information.

This is required for the additional ps feature below.

use_wmi = no

Include the whole path of a process and its arguments in the process list.

full_path = no

···

On Mon, Oct 17, 2016 at 7:12 AM, Wouter wbleeker@gmail.com wrote:

It is not related to your cmk-pusher script.

Can you share the config of check_mk? The regular ini file in which you added the mrpe command?

Also: can you follow up on my question to add the log write action as the first line in your powershell script? It will (probably) show that the powershell script cannot be executed at all. So in that case it is not related to the commands in your scripts.

Verstuurd vanaf mijn iPhone

Op 16 okt. 2016 om 23:21 heeft Pavel Vaysman goam03@gmail.com het volgende geschreven:

i push result with external python program.

https://github.com/qmexnetworks/cmk-pusher-agent

but it’s not powershell problem…

when i run this script manually from powershell it’s working, from debug mode in cmd “check_mk_agent debug” its working , but “check_mk_agent test” - it failed with error…

this mrpe check is working with agent 1.2.6p12…

On Oct 16, 2016 15:06, “Wouter” wbleeker@gmail.com wrote:

So how is your agent started in your situation? You push the results, so it doesn’t run as a service?

Can you get a regular powershell script to write-out any output? When you add something like “add-content full-path-to-logfile.log test_txt” in the first line, is your script running at all?

Verstuurd vanaf mijn iPhone

Op 15 okt. 2016 om 09:05 heeft Pavel Vaysman goam03@gmail.com het volgende geschreven:

I have about 300 server behind firewalls and nat and it’s impossible to query check_mk agent and snmp from central check_mk server.

So i run check_mk agent locally on server and push it’s output to check_mk server.

On Sat, Oct 15, 2016 at 12:05 AM, mlist@libero.it mlist@libero.it wrote:

Pavel,

why are you using a powershell script? in my test environment I added some proliant dl servers running linux + hp agents and it worked out of the box. I did not try windows servers but it should be the same; I just selected agent+snmp

----Messaggio originale----

Da: “Pavel Vaysman” goam03@gmail.com

Data: 14/10/2016 17.27

A: checkmk-en@lists.mathias-kettner.de

Ogg: [Check_mk (english)] windows check_mk_agent 1.2.8p12 with hpacucli.exe

Hi guys,

i try to monitor hp raid with mrpe via check_mk_agent 1.2.8p12 with hpacucli.exe.
I have powershell script that run hpacucli.exe and parsing it’s output.
With check_mk_agent 1.2.6p16 it’s working perfect.
With new check_mk agent 1.2.8pX it’s return “ERROR: Cannot open console”
When i’m run check_mk agent 1.2.8pX with debug argument - it’s working.
Right now , workaround - write hpacucli.exe output to file via windows task scheduler and parse this txt file.

Is it bug or may be i did mistake somewhere?

Source Code

1. $array_config_util = “c:\hp\Bin\hpacucli.exe”
2. try {
3. $exec = & $array_config_util ‘ctrl all show config’
4. #write-host !!!
5. #Write-Output $exec
6. }
7. catch {
8. Write-Host “Problem checking array status)”
9. exit 3
10. }
11. # filter results for lines that talk about drives (physicaldrive, logicaldrive)
12. # and do not end with “OK”:
13. $not_OK = $exec | Where-Object { $_ -like “drive” } | Where-Object { $_ -notlike “OK)” }
14. #write-output “<<<>>>”
15. #write-output $not_OK
16. if ($not_OK.length -lt 1) {
17. Write-Host “Array status appears OK”
18. exit 0
19. }
20. Write-Host “Array status not OK; please check RAID software”
21. exit 2

[Goam03 (Pavel Vaysman)]

i found a problem

Script is not working only in test mode “check_mk_agent test”

cmk-pusher run check_mk_agent with test mode:

command = ConfigSectionMap(“check_mk”)[‘path’]+" test"

With debug, adhoc, file options it’s running just fine.

Is in test mode , agent has to run as in normal mode and only write to stdout?

···

On Mon, Oct 17, 2016 at 10:21 AM, Pavel Vaysman goam03@gmail.com wrote:

Hi,
I added this line to to the script. it’s create “log file” without problem.

this is check_mk.ini file configuration:

Example configuration for Windows agent

[global]

Restrict access to certain IP addresses

If ipv6 is enabled, all listed ipv4 adresses are also accepted as

source adresses in their ipv6-mapped form. I.e. if

192.168.56.0/24 is listed, connections from 0:0:0:0:0:ffff:c0a8:3800/120

are also possible

only_from = 127.0.0.1 192.168.56.0/24 ::1

Change port where the agent is listening ( default 6556 )

port = 6556

Disable ipv6 support. By default, ipv4 is always supported

and ipv6 is enabled additionally if supported by the system.

ipv6 = no

Do only execute programs and scripts with

the following suffixes (in local and plugins).

Per default all extensions except txt and dir

are being executed.

execute = exe bat vbs

Restrict the following configuration variables

in [global] to the following hosts

host = winsrv* zab???ad

Just output certain sections (e.g. for upper hosts)

sections = check_mk winperf

output sections in realtime updates. Please note that not all

sections can be used in realtime updates, the following is the

complete list

realtime_sections = mem df winperf_processor

specifies how long (in seconds) realtime updates are sent to

the last monitoring system that requested an update.

this should be longer than the request frequency (usually

one minute).

Please note that any new request cancels previous realtime

update schedules, so no more than one update is sent per second,

no matter if this timeout is “too high” or how many monitoring

systems are querying the agent.

realtime_timeout = 90

passphrase for encrypted communication. Currently this is only

used in realtime updates

passphrase = secret

Write a logfile for tackling down crashes of the agent

crash_debug = yes

by default the agent flushes the socket for each agent section.

this is mostly for debugging purposes in case the agent is slow

to respond.

Disabling this may (very) slightly reduce network traffic.

section_flush = no

[winperf]

Select counters to extract. The following counters

are needed by checks shipped with check_mk.

counters = 10332:msx_queues

counters = 638:tcp_conn

[logfiles]

# Define textfiles to be monitored, separated by |

textfile = C:\tmp logfiles\message_*.log|D:\log\sample.txt

# Set patterns for defined textfiles

ok = Successfully logged in*

crit = Error in*

warn = Unable to contact*

ignore = Backup * saved

# Define additional textfiles with different patterns

textfile = C:\tmp\memo.udf

# Set patterns for defined textfile

warn = overdue

ok = mail sent

[logwatch]

Testing: output all messages from the eventlogs

sendall = yes

From application log send only critical messages

logfile application = crit

From system log send only warning/critical messages,

but suppress any context messages

logfile system = nocontext warn

Switch all other logfiles off. Default is warn:

send messages of type warn or crit

logfile * = off

[mrpe]

Run classical monitoring plugins. The word before the command

line is the service description for the monitoring. Use backslashes

in Windows-paths.

check = IP_Configuration mrpe\check_ipconfig 1.2.3.4

check = Whatever c:\myplugins\check_whatever -w 10 -c 20

check = HP_Smart_Array c:\Windows\System32\WindowsPowerShell\v1.0\powershell.exe -command “.‘c:\medasys-monitor\mrpe\hp-raid.ps1’; exit($lastexitcode)”

[fileinfo]

path = C:\Programs\Foo*.log

path = M:\Bar Test*.*

[local]

define timeouts for local scripts matching

specific patterns - first match wins

timeout *.vbs = 20

timeout *.bat = 10

timeout * = 30

[plugins]

example: the windows_updates.vbs

plugin is executed asynchronous

and is only updated every 3600 seconds

it may fail (timeout / error) up to 3 times before the last known data is discarded

execution windows_updates.vbs = async

timeout windows_updates.vbs = 120

cache_age windows_updates.vbs = 3600

retry_count windows_udpates.vbs = 3

define timeouts for plugin scripts matching

specific patterns - first match wins

timeout ps_perf.ps1 = 20

timeout *.ps1 = 10

timeout * = 30

When using the Check_MK Inventory plugin, it is a good idea to make the

plugin being executed asynchronous to prevent it from hooking up the

whole agent processing. Additionally it should have a execution timeout.

execution mk_inventory.ps1 = async

timeout mk_inventory.ps1 = 240

[ps]

Experimental: Set to yes to use wmi for retrieving process information.

This is required for the additional ps feature below.

use_wmi = no

Include the whole path of a process and its arguments in the process list.

full_path = no

On Mon, Oct 17, 2016 at 7:12 AM, Wouter wbleeker@gmail.com wrote:

It is not related to your cmk-pusher script.

Can you share the config of check_mk? The regular ini file in which you added the mrpe command?

Also: can you follow up on my question to add the log write action as the first line in your powershell script? It will (probably) show that the powershell script cannot be executed at all. So in that case it is not related to the commands in your scripts.

Verstuurd vanaf mijn iPhone

Op 16 okt. 2016 om 23:21 heeft Pavel Vaysman goam03@gmail.com het volgende geschreven:

i push result with external python program.

https://github.com/qmexnetworks/cmk-pusher-agent

but it’s not powershell problem…

when i run this script manually from powershell it’s working, from debug mode in cmd “check_mk_agent debug” its working , but “check_mk_agent test” - it failed with error…

this mrpe check is working with agent 1.2.6p12…

On Oct 16, 2016 15:06, “Wouter” wbleeker@gmail.com wrote:

So how is your agent started in your situation? You push the results, so it doesn’t run as a service?

Can you get a regular powershell script to write-out any output? When you add something like “add-content full-path-to-logfile.log test_txt” in the first line, is your script running at all?

Verstuurd vanaf mijn iPhone

Op 15 okt. 2016 om 09:05 heeft Pavel Vaysman goam03@gmail.com het volgende geschreven:

I have about 300 server behind firewalls and nat and it’s impossible to query check_mk agent and snmp from central check_mk server.

So i run check_mk agent locally on server and push it’s output to check_mk server.

On Sat, Oct 15, 2016 at 12:05 AM, mlist@libero.it mlist@libero.it wrote:

Pavel,

why are you using a powershell script? in my test environment I added some proliant dl servers running linux + hp agents and it worked out of the box. I did not try windows servers but it should be the same; I just selected agent+snmp

----Messaggio originale----

Da: “Pavel Vaysman” goam03@gmail.com

Data: 14/10/2016 17.27

A: checkmk-en@lists.mathias-kettner.de

Ogg: [Check_mk (english)] windows check_mk_agent 1.2.8p12 with hpacucli.exe

Hi guys,

i try to monitor hp raid with mrpe via check_mk_agent 1.2.8p12 with hpacucli.exe.
I have powershell script that run hpacucli.exe and parsing it’s output.
With check_mk_agent 1.2.6p16 it’s working perfect.
With new check_mk agent 1.2.8pX it’s return “ERROR: Cannot open console”
When i’m run check_mk agent 1.2.8pX with debug argument - it’s working.
Right now , workaround - write hpacucli.exe output to file via windows task scheduler and parse this txt file.

Is it bug or may be i did mistake somewhere?

Source Code

1. $array_config_util = “c:\hp\Bin\hpacucli.exe”
2. try {
3. $exec = & $array_config_util ‘ctrl all show config’
4. #write-host !!!
5. #Write-Output $exec
6. }
7. catch {
8. Write-Host “Problem checking array status)”
9. exit 3
10. }
11. # filter results for lines that talk about drives (physicaldrive, logicaldrive)
12. # and do not end with “OK”:
13. $not_OK = $exec | Where-Object { $_ -like “drive” } | Where-Object { $_ -notlike “OK)” }
14. #write-output “<<<>>>”
15. #write-output $not_OK
16. if ($not_OK.length -lt 1) {
17. Write-Host “Array status appears OK”
18. exit 0
19. }
20. Write-Host “Array status not OK; please check RAID software”
21. exit 2