Monitoring performance of an HP EVA SAN

[gsjodin]

I have been running into some problems with python and wmic. i am thinking of updating my OS from Red Hat 4 enterprise to Red Hat 5 enterprise. Can anyone suggest a recomended OS.

Thanks

Gord

[Alphadog]

Hi Palli,

Scripts and Templates are working fine, just have some problems getting the wmi data of another EVA on a second cacti installation.

Do you know what this error tells me ??

Code: Select all

ERROR: Retrieve result data.
NTSTATUS: NT code 0xc0001010 - NT code 0xc0001010

I ran wmiadap because i didnt even get the Win32_PerfRawData out of the machine but i dont get any eva data -.-

edit:
could you post the link to your thread on ITRC ??

[Alphadog]

ok solved my Problem by reinstalling Command view ...
but now i dont get any new graphs ...

its hairpulling ...

[palli]

I have been running into some problems with python and wmic. i am thinking of updating my OS from Red Hat 4 enterprise to Red Hat 5 enterprise. Can anyone suggest a recomended OS.

Thanks

Gord

Gord: These templates work very well with Redhat Enterprise Linux 5.x

they are not going to work well on 4.x except with some hacking.

Also remember you will need cacti version at least 0.8.7e which is available in EPEL.

[palli]

ok solved my Problem by reinstalling Command view ...
but now i dont get any new graphs ...

its hairpulling ...

Yes i agree, troubleshooting, evaperf+wmi+python+cacti can all be frustrating, especially in this case when the problem can be with in any single component of these.

I have been installing these templates in quite a few environments, these are the troubleshooting tricks i use the most often:

1) Check if WMI is working at all:

Code: Select all

 /usr/local/bin/wmic //HOSTNAME --user administrator "select Name from Win32_ComputerSystem"

If it returns the name of the command-view server i know WMI is working. Then i check if evaperf extensions of are working (they are very fragile and a common point of failure).

2) Check if WMI is working with evaperf:

Code: Select all

/usr/local/bin/wmic //HOSTNAME --user administrator "select Name from Win32_PerfRawData_EVAPMEXT_HPEVAHostConnection"

Now, if this fails (for example with 0xc0001010) it means that evaperf has crashed for some reason. It might help to:

* Run "wmiadap /f" on the windows machine
* Restarting HP evaperf service in windows services
* Restarting Windows Management Interface in windows services

Results may vary, if the first wmic works but the second fails i urge you to contact HP support and let them know that evaperf is not stable enough for your needs!

If both wmic commands work

* Make sure the latter command gave you human readable hostnames, if you got bunch of numbers you will have to refresh evaperf friendly-names in order to get human readable friendlynames in cacti

* If wmic works but you are unable to query devices in cacti, it might help to check out /var/log/cacti/query_eva.log it is quite verbose but errors but you might catch errors from wmic there.

Good luck![/list]

[Alphadog]

Hey guys,

my WMI does now work quite well except that its crashing sometimes.

I wrote a simple batch file, that stops the WMI service imports the *.reg file, which sets the DWORD

Code: Select all

"HKLM\SYSTEM\ControlSet001\Services\EVAPMEXT\Performance\DisablePerformanceCounters"

to "0", starts the wmi service and runs

Code: Select all

wmiadap /F

This one i put under Planned Tasks an let it run every hour.

If anyone wants the batch script its in my attachment

Greez

Alpha

[palli]

Hey guys,

my WMI does now work quite well except that its crashing sometimes.

Nice! My experience is that WMI is quite stable, but it is the evaperf extensions to wmi that keep crashing. If you have a support contract with HP, dont hesitate to contact them and let them know evaperf is not working as you would like it.

[Alphadog]

wmi sucks -.-

New Machine => New Installation => 2 new EVA's (4400) => no response from wmi

Win32_PerfRawData gets me data
Win32_PerfRawData_EVAPMEXT_HPEVAHostPortStatistics dont gets me anything ...

DisablePerformanceCounters is "0"
WMI Service is running

When restarting it I can get 1 dataset and then all other outputs are empty

anyone who could help me?

[palli]

wmi sucks -.-

New Machine => New Installation => 2 new EVA's (4400) => no response from wmi

Win32_PerfRawData gets me data
Win32_PerfRawData_EVAPMEXT_HPEVAHostPortStatistics dont gets me anything ...

DisablePerformanceCounters is "0"
WMI Service is running

When restarting it I can get 1 dataset and then all other outputs are empty

anyone who could help me?

You can try these (in the following order)
* Restart HP Evaperf service
* Run winmgmt /resyncperf
* Run wmiadap /f
* Restart Windows Management Infrastructure service
* Give HP a call, ask them to give you a hand

[thenrique]

I have a problem in graphs of throughput, they are not being generated in debug mode the following error is displayed:

Code: Select all

"RRDTool Says:
ERROR: invalid rpn expression in: a, 1024, * "

Does anyone have any idea what could be wrong?

I also created a list of friendly names, but it is not being consulted by wmic, what can be wrong?

[gandalf]

I have a problem in graphs of throughput, they are not being generated in debug mode the following error is displayed:

Code: Select all

"RRDTool Says:
ERROR: invalid rpn expression in: a, 1024, * "

Does anyone have any idea what could be wrong?

I also created a list of friendly names, but it is not being consulted by wmic, what can be wrong?

I'd say: omit the whitespaces
R.

[thenrique]

I have a problem in graphs of throughput, they are not being generated in debug mode the following error is displayed:

Code: Select all

"RRDTool Says:
ERROR: invalid rpn expression in: a, 1024, * "

Does anyone have any idea what could be wrong?

I also created a list of friendly names, but it is not being consulted by wmic, what can be wrong?

I'd say: omit the whitespaces
R.

Thanks, the problem whit friendlynames is resolved restarting the wmi service and hp eva data service collector.

But there no whitespaces in the command for rrd create:

Code: Select all

   	 RRDTool Command:

/usr/bin/rrdtool graph - \
--imgformat=PNG \
--start=-86400 \
--end=-300 \
--title="HostPort EVA8400_ECE:J02 bytes/sec" \
--base=1024 \
--height=120 \
--width=500 \
--upper-limit=1048576 \
--lower-limit=0 \
--vertical-label="bytes/sec" \
--slope-mode \
--font TITLE:12: \
--font AXIS:8: \
--font LEGEND:10: \
--font UNIT:8: \
CDEF:cdefa=a,1024,* \
LINE1:cdefa#7CB3F1FF:"Read Bytes"  \
AREA:cdefa#7CB3F17F:""  \
GPRINT:cdefa:LAST:"           Current\:%8.2lf %s"  \
GPRINT:cdefa:AVERAGE:"Average\:%8.2lf %s"  \
GPRINT:cdefa:MAX:"Maximum\:%8.2lf %s\n"  \
LINE2:cdefa#FF0000FF:"Write Bytes"  \
GPRINT:cdefa:LAST:"Current\:%8.2lf %s"  \
GPRINT:cdefa:AVERAGE:"Average\:%8.2lf %s"  \
GPRINT:cdefa:MAX:"Maximum\:%8.2lf %s\n" 

RRDTool Says:

ERROR: invalid rpn expression in: a,1024,*

[/quote]

[Alphadog]

Ok for everyone who has problems with a crashing Wmi service, here's a little howto:

There are 4 reasons why your cacti doesnt graph or doesnt even get some data from the command view host:

Reason One:

wmic command gets you this :

Code: Select all

ERROR: Retrieve result data. 
NTSTATUS: NT code 0xc0001010 - NT code 0xc0001010 

Solution:

Check if there is a "HP WMI Storage Providers" Service
If yes check if there are the "HP EVA" Counters in Perfmon
If yes try Stopping the WMI Service (the Windows one) restart the "HP WMI" and start the windows WMI Service.

Case Two:

Services are existent, restarted and running but NO COUNTERS in Perfmon

Solution:

Go to your Registry under

Code: Select all

[HKLM/SYSTEM/ControlSet001/Services/EVAPMEXT/Performance/]

And Search for a DWORD Called "Disable Performance Counters"
Check if this is set to "0" no matter if Hex or Decimal 0=0 xD
Restart the Services as explained above
Run

Code: Select all

wmiadap /F

in cmd


Case Three

Now it gets Tricky

Servcices are existent, restarted and Running, "Disable Performance Counters" is set to "0" but counters dont appear in Perfmon

Solution:
Follow the above Steps, run

Code: Select all

wmiadap /F

and so on
AND
if you have done this:

Run

Code: Select all

lodctr /R

in your cmd

Case Four

Followed all the steps above, wmic get some data BUT Cacti doesn't even Poll the new graphs created

Solution

Try Changing the Poller from Spine to cmd.php because Spine sometimes gets a "Buffer Overflow" or something like that doesnt matter whic error, it just crashes and doesnt poll the new graphs







###### EDIT ######
Please do me a favour and edit all you Logfile Directories and place it into /var/log/Cacti/
I have some some servers, which have now a directory with 3 GB of Logs, and in /var/log the logs are automaticly archived, which increases performance when viewing tried to open a 400MB log and had to wait about half an hour ^^

###### EDIT ######




If there's anything else not working, contact me either in this Fred or PM me I'll see what I can do for you

Greez from sunny Bavaria

Alphadog[/b]

[jshroyer]

I have installed the templates, as instructed i believe, however the python script (query_wmi.py) does not seem to be returning any data. I can pull data via wmic, when I try to run the command as given in the script debug output, it says:

Traceback (most recent call last):
File "/cacti/scripts/query_wmi.py", line 10, in ?
from subprocess import Popen,PIPE,STDOUT
ImportError: No module named subprocess


debug output from cacti:
+ Running data query [14].
+ Found type = '4 '[script query].
+ Found data query XML file at '/cacti/resource/script_queries/eva_storagecontroller.xml'
+ XML file parsed ok.
+ Executing script for list of indexes 'python /cacti/scripts/query_wmi.py --host HOSTNAME --authentication-file /etc/cacti/HOSTNAME.pw --class Win32_PerfRawData_EVAPMEXT_HPEVAStorageController index'
+ Executing script query 'python /cacti/scripts/query_wmi.py --host HOSTNAME --authentication-file /etc/cacti/HOSTNAME.pw --class Win32_PerfRawData_EVAPMEXT_HPEVAStorageController query Name'
+ Found data query XML file at '/cacti/resource/script_queries/eva_storagecontroller.xml'
+ Found data query XML file at '/cacti/resource/script_queries/eva_storagecontroller.xml'
+ Found data query XML file at '/cacti/resource/script_queries/eva_storagecontroller.xml'

Any help is appreciated.

[willdej]

I downloaded and installed the EVA Commandview template
When I try to do a verbose it seems like it works but under status I get Success [ 0 Items, 0 Rows]

When I try to run /usr/local/bin/wmic //192.168.1.11 --user administrator
I get a Floating Point Exception error.
I assume that possible I have wmi not working correct?
I am running 0.8.7e on a SLES 10 SP2 server