SNMP-Informant-STD Memory Usage (Windows)

Templates, scripts for templates, scripts and requests for templates.

Moderators: Developers, Moderators

ajeskey
Cacti User
Posts: 93
Joined: Mon Mar 28, 2005 4:31 pm
Location: Boston, MA

SNMP-Informant-STD Memory Usage (Windows)

Post by ajeskey »

This uses the SNMP-Informants Standard MIB Add-On to graph Memory stats on Windows.

You must have SNMP-Informant Standard MIB Add-On install on the Windows host you are polling in order for this to work. If you are not familiar with SNMP-Informant, it is a free download, and is extremely stable, since it does not add any services to the machine.

I have tested this on windows 2000 and windows 2003 systems and seems to be accurate.

Note
-----
There is a custom CDEF included in this that requiers the order remain the same in the graph template, if you change it it will break.


What it Graphs
--------------------
* Installed Memory (Static)
* Available Memory (Current and Average)
* Used Memory (Current and Average)
* Cache Resident Memory (Current and Average)
* Page Faults (Current and Average)
Attachments
SNMP-Informant-Memory.JPG
SNMP-Informant-Memory.JPG (41.72 KiB) Viewed 32904 times
cacti_graph_template_snmp_informantstd_memory.xml
(25.92 KiB) Downloaded 4425 times
- Aaron Jeskey
chadw
Cacti User
Posts: 52
Joined: Thu Jul 17, 2003 7:20 am

Post by chadw »

Thank you for the snmp-informant templates! I was having trouble getting these parameters with the default Windows templates.

I'm having some trouble with this graph though. The installed memory is nan, as is used memory. I'm guessing the used memory parameter is simply installed memory - available memory, so it also becomes nan.

I can see the OID for installed mem from snmp:

# snmpwalk -v 1 -c public x.x.x.x .1.3.6.1.2.1.25.2.2.0
HOST-RESOURCES-MIB::hrMemorySize.0 = INTEGER: 1048112 KBytes

The polled machine is Windows 2000.

Any idea why this isn't working?
Attachments
untitled.gif
untitled.gif (7.2 KiB) Viewed 32839 times
ajeskey
Cacti User
Posts: 93
Joined: Mon Mar 28, 2005 4:31 pm
Location: Boston, MA

Post by ajeskey »

What happens when you poll .1.3.6.1.2.1.25.2.2.0 with SNMPv2? I am getting the correct info from Host-Resourves-V2, I never tested with v1

If you want to force this graph use V1 go to the Data Template "SNMP - Informant-STD - Memory Installed", at the bottom you'll see "SNMP Version (1, 2, or 3)" but a 1 in there and see if it works.

You may end up with a crazy number, if that is the case try turning the applied CDEF for the Installed memory GPRINT to "None" (its set to "Multiply by 1024" now)

You are correct about the Used Memory; it is derived from Installed - Available.

Let me know if any of that helps
- Aaron Jeskey
chadw
Cacti User
Posts: 52
Joined: Thu Jul 17, 2003 7:20 am

Post by chadw »

I don't think changing the snmp version helps:

# snmpwalk -v 2c -c public x.x.x.x .1.3.6.1.2.1.25.2.2.0
HOST-RESOURCES-MIB::hrMemorySize.0 = INTEGER: 1048112 KBytes

# snmpwalk -v 1 -c public x.x.x.x .1.3.6.1.2.1.25.2.2.0
HOST-RESOURCES-MIB::hrMemorySize.0 = INTEGER: 1048112 KBytes

Could the problem be the "KBytes" that the snmpwalk returns? It's not a number.

If I dump the rra file with rrdtool dump, all of the records have NaN.
ajeskey
Cacti User
Posts: 93
Joined: Mon Mar 28, 2005 4:31 pm
Location: Boston, MA

Post by ajeskey »

no since its a returned integer, your snmpwalk tool is just transforming it to kbytes.

First try turning off the CDEF for the Installed memory under the graph template. If that doesnt help, try pulling all the data sources data templates and graph referances out, reimport it.

Are you using the default SNMPD that comes with windows?
- Aaron Jeskey
chadw
Cacti User
Posts: 52
Joined: Thu Jul 17, 2003 7:20 am

Post by chadw »

I turned the CDEFs off and it didn't make a difference. Yes, the windows servers are using the default snmpd. I'm running cacti on Redhat ES 3.

I think the problem is with getting the data into the rra file. I see only NaN in my rrdtool dumps.

I'm using rrdtool 1.0.49. Should I try a 1.2 release? I'm also using cactid 0.8.6d on cacti-0.8.6e.
ajeskey
Cacti User
Posts: 93
Joined: Mon Mar 28, 2005 4:31 pm
Location: Boston, MA

Post by ajeskey »

I am on "c" and runing 1.0.49

I would tend to belive versions mis match is your problem, since both things that don't show up have CDEFs applied to them; and CDEFs are an RRDtool function. try getting your RRD. I know that when I tried to get "d" working with my current version of RRDtool nothing workd.

I am fairly sure that you should be runnign 1.2 if you are on "d"
- Aaron Jeskey
ajeskey
Cacti User
Posts: 93
Joined: Mon Mar 28, 2005 4:31 pm
Location: Boston, MA

Post by ajeskey »

I am curious to hear if anyone else is having issues with this. I see a number of people have downloaded it... anyone?
- Aaron Jeskey
chadw
Cacti User
Posts: 52
Joined: Thu Jul 17, 2003 7:20 am

Post by chadw »

I've upgraded to rrdtool 1.2.6 (the only RPM I could install without upgrading perl).

The problem is a little different now. Now instead of NaN, I get 0.00.
Attachments
untitled.gif
untitled.gif (11.29 KiB) Viewed 32736 times
ajeskey
Cacti User
Posts: 93
Joined: Mon Mar 28, 2005 4:31 pm
Location: Boston, MA

Post by ajeskey »

try turnign off the CDEF now for installed memory just to see what happens. its a good sign that used shows up as a negitive number now, it means that the CDEF is actuayl working now. So its an issue with the installed memory being read from the machine (or how cacti is handeling that)
- Aaron Jeskey
chadw
Cacti User
Posts: 52
Joined: Thu Jul 17, 2003 7:20 am

Post by chadw »

I turned the CDEF off, it made no difference. Here's some output from my .rra:
# rrdtool dump xxx_meminforminstalled_2147.rrd | grep row | grep -v NaN
<pdp_per_row> 1 </pdp_per_row> <!-- 300 seconds -->
<!-- 2005-06-23 13:20:00 NDT / 1119541800 --> <row><v> 0.0000000000e+00 </v></row>
<pdp_per_row> 6 </pdp_per_row> <!-- 1800 seconds -->
Notice that it only updated the RRD once, when it was created, and it was updated with a 0.00000.

Here's the memAvailable file, it has many updates:
# rrdtool dump xxx_meminformavailable_2145.rrd | grep row | grep -v NaN
<pdp_per_row> 1 </pdp_per_row> <!-- 300 seconds -->
<!-- 2005-06-23 13:20:00 NDT / 1119541800 --> <row><v> 7.5015140818e+08 </v></row>
<!-- 2005-06-23 13:25:00 NDT / 1119542100 --> <row><v> 7.5792793600e+08 </v></row>
<!-- 2005-06-23 13:30:00 NDT / 1119542400 --> <row><v> 7.5661721600e+08 </v></row>
<!-- 2005-06-23 13:35:00 NDT / 1119542700 --> <row><v> 7.5102617600e+08 </v></row>
<!-- 2005-06-23 13:40:00 NDT / 1119543000 --> <row><v> 7.5650468523e+08 </v></row>
<!-- 2005-06-23 13:45:00 NDT / 1119543300 --> <row><v> 7.5680877227e+08 </v></row>
<!-- 2005-06-23 13:50:00 NDT / 1119543600 --> <row><v> 7.5575909035e+08 </v></row>
<!-- 2005-06-23 13:55:00 NDT / 1119543900 --> <row><v> 7.5739136000e+08 </v></row>
<pdp_per_row> 6 </pdp_per_row> <!-- 1800 seconds -->
If I get the value via snmpwalk, I get a response that's not zero:
# snmpwalk -v 2c -c public x.x.x.x .1.3.6.1.2.1.25.2.2.0
HOST-RESOURCES-MIB::hrMemorySize.0 = INTEGER: 1179120 KBytes
Thanks,

Chad
chadw
Cacti User
Posts: 52
Joined: Thu Jul 17, 2003 7:20 am

Post by chadw »

Here's the cacti.log section where it puts the value into poller_output:
06/23/2005 02:30:25 PM - CACTID: Poller[0] Host[136] SNMP: v2: x.x.x.x, dsname: meminformInstalled, oid: .1.3.6.1.2.1.25.2.2.0, value: 1048112 KBytes
06/23/2005 02:30:25 PM - CACTID: Poller[0] DEBUG: SQLCMD: insert into poller_output (local_data_id,rrd_name,time,output) values (2176,'meminformInstalled','2005-06-23 14:30:18','1048112 KBytes')
Here's the output from cacti.log when the .rra gets created:
06/23/2005 02:31:23 PM - POLLER: Poller[0] CACTI2RRD: /usr/bin/rrdtool create /usr/local/cacti-0.8.6e/rra/nltechsupport_meminforminstalled_2176.rrd --step 300 DS:meminformInstalled:GAUGE:600:0:U RRA:AVERAGE:0.5:1:600 RRA:AVERAGE:0.5:6:700 RRA:AVERAGE:0.5:24:775 RRA:AVERAGE:0.5:288:797 RRA:MIN:0.5:1:600 RRA:MIN:0.5:6:700 RRA:MIN:0.5:24:775 RRA:MIN:0.5:288:797 RRA:MAX:0.5:1:600 RRA:MAX:0.5:6:700 RRA:MAX:0.5:24:775 RRA:MAX:0.5:288:797 RRA:LAST:0.5:1:600 RRA:LAST:0.5:6:700 RRA:LAST:0.5:24:775 RRA:LAST:0.5:288:797

06/23/2005 02:31:23 PM - POLLER: Poller[0] CACTI2RRD: /usr/bin/rrdtool update /usr/local/cacti-0.8.6e/rra/nltechsupport_meminforminstalled_2176.rrd --template meminformInstalled N:U
Somewhere between the poll and updating poller_output, the value gets changed from 1048112 to U.
chadw
Cacti User
Posts: 52
Joined: Thu Jul 17, 2003 7:20 am

Post by chadw »

Upon further checking, it looks like the problem is this line:
06/23/2005 02:30:25 PM - CACTID: Poller[0] DEBUG: SQLCMD: insert into poller_output (local_data_id,rrd_name,time,output) values (2176,'meminformInstalled','2005-06-23 14:30:18','1048112 KBytes')
I manually inserted a record into poller_output while it was populating from cactid without the "KBytes" string and the .rra file updated correctly:
insert into poller_output (local_data_id,rrd_name,time,output) values (2176,'meminformInstalled','2005-06-23 14:30:18','1048112')
Now, how do I prevent it from happening?
ajeskey
Cacti User
Posts: 93
Joined: Mon Mar 28, 2005 4:31 pm
Location: Boston, MA

Post by ajeskey »

you've got me on this one. You may want to start a thread in the general help forums and point them this way.

Glad to see that its working. Does it look good / what you wanted?

I think the next thing I should do with this is get it to report in bytes... but I've got a few others under my belts ATM, we'll see.
- Aaron Jeskey
chadw
Cacti User
Posts: 52
Joined: Thu Jul 17, 2003 7:20 am

Post by chadw »

I posted a message in the general forum and The Witness told me to upgrade cactid to 0.8.6e, I'm currently using 0.8.6d. He said the e version would strip the KBytes string off the value.

Thanks again for your help.
Post Reply

Who is online

Users browsing this forum: No registered users and 7 guests