Windows servers uptime values

Previous Topic Next Topic
 
classic Classic list List threaded Threaded
13 messages Options
Reply | Threaded
Open this post in threaded view
|

Windows servers uptime values

James Zuelow
I seem to be getting crazy large values for uptime on my windows servers
with 1.5.93.

With my 1.2.9 installation I had edited datacollection-config.xml to
poll .1.3.6.1.2.1.1.3.0 for uptime values.  Those values agreed with
perfmon and systeminfo.

The 1.5.93 installation is polling OID 1.3.6.1.2.1.25.1.1, which results
in values about (but not quite) twice as large.

The second OID is correct for Linux and AIX machines on our network, but
definitely not for Server 2003 or Server 2008.

With 1.2.9 it was easy to drop in the Windows uptime graph because 1.2.9
was not collecting any of the net-snmp OIDs for Windows servers.  So I
could name the rrd the same as the net-snmp rrd, and the graph would
magically appear.

What would be the best way of replacing that OID for just windows
machines?

I am thinking that I have to pull the net-snmp include out of the
Windows section, and then copy-paste the guts of the section back in
with the OID that I want to change edited.

Is there an easier way? (Or better yet, am I the only one seeing strange
uptime numbers on my Windows servers?)

James Zuelow....................CBJ MIS (907)586-0236
Network Specialist...Registered Linux User No. 186591

-------------------------------------------------------------------------
This SF.Net email is sponsored by the Moblin Your Move Developer's challenge
Build the coolest Linux based applications with Moblin SDK & win great prizes
Grand prize is a trip for two to an Open Source event anywhere in the world
http://moblin-contest.org/redirect.php?banner_id=100&url=/
_______________________________________________
Please read the OpenNMS Mailing List FAQ:
http://www.opennms.org/index.php/Mailing_List_FAQ

opennms-discuss mailing list

To *unsubscribe* or change your subscription options, see the bottom of this page:
https://lists.sourceforge.net/lists/listinfo/opennms-discuss
Reply | Threaded
Open this post in threaded view
|

Re: Windows servers uptime values

(private) HKS
1.3.6.1.2.1.25.1.1 is reported by Windows in milliseconds rather than
centisecond (as Net-SNMP reports it). I've gotten around this in a
mixed environment by removing mib2-host-resources-system from the
Windows system and replacing it with the windows-host group:

      <group name="windows-host" ifType="ignore">
        <mibObj oid=".1.3.6.1.2.1.25.3.3.1.2" instance="1"
alias="cpuPercentBusy" type="integer" />
        <!-- Memory in KBytes -->
        <mibObj oid=".1.3.6.1.2.1.25.2.2" instance="0"
alias="memorySize" type="integer" />
        <mibObj oid=".1.3.6.1.2.1.25.1.1" instance="0"
alias="winSystemUptime" type="timeticks" />
        <mibObj oid=".1.3.6.1.2.1.25.1.5" instance="0"
alias="hrSystemNumUsers" type="gauge" />
        <mibObj oid=".1.3.6.1.2.1.25.1.6" instance="0"
alias="hrSystemProcesses" type="gauge" />
      </group>

Then add the following graph to snmp-graph.properties:

report.custom.windows.winSystemUptime.name=System Uptime (Windows)
report.custom.windows.winSystemUptime.columns=winSystemUptime
report.custom.windows.winSystemUptime.type=nodeSnmp
report.custom.windows.winSystemUptime.command=--title="System Uptime" \
 --vertical-label Days \
 DEF:time={rrd1}:winSystemUptime:AVERAGE \
 CDEF:days=time,86400000,/ \
 LINE2:days#0000ff:"System Uptime (Days)" \
 GPRINT:days:AVERAGE:"Avg  \\: %8.1lf %s" \
 GPRINT:days:MIN:"Min  \\: %8.1lf %s" \
 GPRINT:days:MAX:"Max  \\: %8.1lf %s\\n"


Be sure to remove mib2-host-resources-system from the Windows SNMP
system or you'll have two graphs, one of which is a tad confusing.

-HKS

On Mon, Sep 8, 2008 at 2:19 PM, James Zuelow
<[hidden email]> wrote:

> I seem to be getting crazy large values for uptime on my windows servers
> with 1.5.93.
>
> With my 1.2.9 installation I had edited datacollection-config.xml to
> poll .1.3.6.1.2.1.1.3.0 for uptime values.  Those values agreed with
> perfmon and systeminfo.
>
> The 1.5.93 installation is polling OID 1.3.6.1.2.1.25.1.1, which results
> in values about (but not quite) twice as large.
>
> The second OID is correct for Linux and AIX machines on our network, but
> definitely not for Server 2003 or Server 2008.
>
> With 1.2.9 it was easy to drop in the Windows uptime graph because 1.2.9
> was not collecting any of the net-snmp OIDs for Windows servers.  So I
> could name the rrd the same as the net-snmp rrd, and the graph would
> magically appear.
>
> What would be the best way of replacing that OID for just windows
> machines?
>
> I am thinking that I have to pull the net-snmp include out of the
> Windows section, and then copy-paste the guts of the section back in
> with the OID that I want to change edited.
>
> Is there an easier way? (Or better yet, am I the only one seeing strange
> uptime numbers on my Windows servers?)
>
> James Zuelow....................CBJ MIS (907)586-0236
> Network Specialist...Registered Linux User No. 186591
>
> -------------------------------------------------------------------------
> This SF.Net email is sponsored by the Moblin Your Move Developer's challenge
> Build the coolest Linux based applications with Moblin SDK & win great prizes
> Grand prize is a trip for two to an Open Source event anywhere in the world
> http://moblin-contest.org/redirect.php?banner_id=100&url=/
> _______________________________________________
> Please read the OpenNMS Mailing List FAQ:
> http://www.opennms.org/index.php/Mailing_List_FAQ
>
> opennms-discuss mailing list
>
> To *unsubscribe* or change your subscription options, see the bottom of this page:
> https://lists.sourceforge.net/lists/listinfo/opennms-discuss
>

-------------------------------------------------------------------------
This SF.Net email is sponsored by the Moblin Your Move Developer's challenge
Build the coolest Linux based applications with Moblin SDK & win great prizes
Grand prize is a trip for two to an Open Source event anywhere in the world
http://moblin-contest.org/redirect.php?banner_id=100&url=/
_______________________________________________
Please read the OpenNMS Mailing List FAQ:
http://www.opennms.org/index.php/Mailing_List_FAQ

opennms-discuss mailing list

To *unsubscribe* or change your subscription options, see the bottom of this page:
https://lists.sourceforge.net/lists/listinfo/opennms-discuss
Reply | Threaded
Open this post in threaded view
|

Re: Windows servers uptime values

Jeff Gehlbach
In reply to this post by James Zuelow
On Sep 8, 2008, at 2:19 PM, James Zuelow wrote:
> I seem to be getting crazy large values for uptime on my windows  
> servers
> with 1.5.93.
 >
> With my 1.2.9 installation I had edited datacollection-config.xml to
> poll .1.3.6.1.2.1.1.3.0 for uptime values.  Those values agreed with
> perfmon and systeminfo.

1.3.6.1.2.1.1.3 is sysUpTime.  It's defined in RFC1213 as:

  "The time (in hundredths of a second) since the network management  
portion of the system was last re-initialized."

That means that sysUpTime is the number of centiseconds since the SNMP  
agent on your Windows boxes was last started.  If you "net stop snmp ;  
net start snmp", then sysUpTime will go back to zero.  That means that  
sysUpTime is totally unreliable as an indicator of actual system uptime.

> The 1.5.93 installation is polling OID 1.3.6.1.2.1.25.1.1, which  
> results
> in values about (but not quite) twice as large.

1.3.6.1.2.1.25.1.1 is hrSystemUptime.  It's defined in RFC1514 (aka  
HOST-RESOURCES-MIB) as:

"The amount of time since this host was last initialized.  Note that  
this is different from sysUpTime in the SNMPv2-MIB [RFC1907] because  
sysUpTime is the uptime of the network management portion of the  
system."

The syntax given for hrSystemUptime is "TimeTicks", which means  
centiseconds, so the units are the same as for sysUpTime.

Any appearance of a correlation between the values of sysUpTime and  
hrSystemUptime is strictly coincidental, assuming the author of the  
SNMP agent from which you queried those values implemented both  
objects properly.

> The second OID is correct for Linux and AIX machines on our network,  
> but
> definitely not for Server 2003 or Server 2008.

Most modern SNMP agents support both sysUpTime and hrSystemUptime.

> With 1.2.9 it was easy to drop in the Windows uptime graph because  
> 1.2.9
> was not collecting any of the net-snmp OIDs for Windows servers.  So I
> could name the rrd the same as the net-snmp rrd, and the graph would
> magically appear.

Uhh, hrSystemUptime is not a "net-snmp OID".  Also, neither 1.2.9 nor  
1.5.93 includes sysUpTime in its default datacollection-config.xml file.

> What would be the best way of replacing that OID for just windows
> machines?
>
> I am thinking that I have to pull the net-snmp include out of the
> Windows section, and then copy-paste the guts of the section back in
> with the OID that I want to change edited.
>
> Is there an easier way? (Or better yet, am I the only one seeing  
> strange
> uptime numbers on my Windows servers?)

I'm afraid you're making less sense the further down I read, which  
probably means that you started with a faulty assumption.  Perhaps you  
can provide actual examples of the values you're seeing for each of  
sysUpTime and hrSystemUptime, alongside the general kinds of values  
you're expecting?

-jeff

-------------------------------------------------------------------------
This SF.Net email is sponsored by the Moblin Your Move Developer's challenge
Build the coolest Linux based applications with Moblin SDK & win great prizes
Grand prize is a trip for two to an Open Source event anywhere in the world
http://moblin-contest.org/redirect.php?banner_id=100&url=/
_______________________________________________
Please read the OpenNMS Mailing List FAQ:
http://www.opennms.org/index.php/Mailing_List_FAQ

opennms-discuss mailing list

To *unsubscribe* or change your subscription options, see the bottom of this page:
https://lists.sourceforge.net/lists/listinfo/opennms-discuss
Reply | Threaded
Open this post in threaded view
|

Re: Windows servers uptime values

Jeff Gehlbach
In reply to this post by (private) HKS
On Sep 8, 2008, at 2:30 PM, (private) HKS wrote:

> 1.3.6.1.2.1.25.1.1 is reported by Windows in milliseconds rather than
> centisecond (as Net-SNMP reports it).

Very interesting, there's my new learning for the day.

If this is true, then I would say it's actually a bug in the Windows  
SNMP Service rather than Windows itself.  And make no mistake, it's a  
bug -- hrSystemUptime is specified as TimeTicks, which has always  
meant centiseconds.

> Be sure to remove mib2-host-resources-system from the Windows SNMP
> system or you'll have two graphs, one of which is a tad confusing.

Or set the "suppress" sub-property on the  
custom.windows.winSystemUptime report to "netsnmp.hrSystemUptime".

-jeff

-------------------------------------------------------------------------
This SF.Net email is sponsored by the Moblin Your Move Developer's challenge
Build the coolest Linux based applications with Moblin SDK & win great prizes
Grand prize is a trip for two to an Open Source event anywhere in the world
http://moblin-contest.org/redirect.php?banner_id=100&url=/
_______________________________________________
Please read the OpenNMS Mailing List FAQ:
http://www.opennms.org/index.php/Mailing_List_FAQ

opennms-discuss mailing list

To *unsubscribe* or change your subscription options, see the bottom of this page:
https://lists.sourceforge.net/lists/listinfo/opennms-discuss
Reply | Threaded
Open this post in threaded view
|

Re: Windows servers uptime values

James Zuelow
In reply to this post by Jeff Gehlbach
OK.  The values won't be identical as I'm doing them one after another.
But they should be within a few minutes.

One of my windows systems has been up for just about 60 days.  The
easiest way to get that into an e-mail is systeminfo output:

Microsoft Windows [Version 5.2.3790]
(C) Copyright 1985-2003 Microsoft Corp.

C:\Documents and Settings\james_zuelow>systeminfo

Host Name:                 CITY-DC1-W3S
OS Name:                   Microsoft(R) Windows(R) Server 2003, Standard
Edition

OS Version:                5.2.3790 Service Pack 2 Build 3790
OS Manufacturer:           Microsoft Corporation
OS Configuration:          Primary Domain Controller
OS Build Type:             Multiprocessor Free
Registered Owner:          CBJ
Registered Organization:   CBJ
Product ID:                69712-640-5317685-45192
Original Install Date:     4/27/2005, 3:40:50 PM
System Up Time:            59 Days, 20 Hours, 15 Minutes, 22 Seconds
System Manufacturer:       IBM

(snip)

Here are the two snmpwalk values:

DISMAN-EVENT-MIB::sysUpTimeInstance = Timeticks: (517059274) 59 days,
20:16:32.74

HOST-RESOURCES-MIB::hrSystemUptime.0 = Timeticks: (875770329) 101 days,
8:41:43.29

Now regardless of what the RFC's say, I **know** that this machine was
last rebooted about 2 months ago.  Not three and a half months ago.

I admit that I may be working from a faulty assumption, but I still want
to graph the first value as it is the accurate value for all of my
Windows machines.

I do not want to graph the second value, as it does not seem to have any
basis in reality for my Windows machines and gets further out of sync
with time.  It *is* accurate for my *nix machines however, so I only
want to replace the polled OID for Windows machines.  That is why I
wanted to edit datacollection-config.xml.

However I also want to stay as close to a stock install as possible, so
was wondering if there is an easier way.

James Zuelow....................CBJ MIS (907)586-0236
Network Specialist...Registered Linux User No. 186591

-----Original Message-----
From: [hidden email]
[mailto:[hidden email]] On Behalf Of Jeff
Gehlbach
Sent: Monday, September 08, 2008 10:42 AM
To: General OpenNMS Discussion
Subject: Re: [opennms-discuss] Windows servers uptime values

On Sep 8, 2008, at 2:19 PM, James Zuelow wrote:
> I seem to be getting crazy large values for uptime on my windows  
> servers
> with 1.5.93.
 >
> With my 1.2.9 installation I had edited datacollection-config.xml to
> poll .1.3.6.1.2.1.1.3.0 for uptime values.  Those values agreed with
> perfmon and systeminfo.

1.3.6.1.2.1.1.3 is sysUpTime.  It's defined in RFC1213 as:

  "The time (in hundredths of a second) since the network management  
portion of the system was last re-initialized."

That means that sysUpTime is the number of centiseconds since the SNMP  
agent on your Windows boxes was last started.  If you "net stop snmp ;  
net start snmp", then sysUpTime will go back to zero.  That means that  
sysUpTime is totally unreliable as an indicator of actual system uptime.

> The 1.5.93 installation is polling OID 1.3.6.1.2.1.25.1.1, which  
> results
> in values about (but not quite) twice as large.

1.3.6.1.2.1.25.1.1 is hrSystemUptime.  It's defined in RFC1514 (aka  
HOST-RESOURCES-MIB) as:

"The amount of time since this host was last initialized.  Note that  
this is different from sysUpTime in the SNMPv2-MIB [RFC1907] because  
sysUpTime is the uptime of the network management portion of the  
system."

The syntax given for hrSystemUptime is "TimeTicks", which means  
centiseconds, so the units are the same as for sysUpTime.

Any appearance of a correlation between the values of sysUpTime and  
hrSystemUptime is strictly coincidental, assuming the author of the  
SNMP agent from which you queried those values implemented both  
objects properly.

> The second OID is correct for Linux and AIX machines on our network,  
> but
> definitely not for Server 2003 or Server 2008.

Most modern SNMP agents support both sysUpTime and hrSystemUptime.

> With 1.2.9 it was easy to drop in the Windows uptime graph because  
> 1.2.9
> was not collecting any of the net-snmp OIDs for Windows servers.  So I
> could name the rrd the same as the net-snmp rrd, and the graph would
> magically appear.

Uhh, hrSystemUptime is not a "net-snmp OID".  Also, neither 1.2.9 nor  
1.5.93 includes sysUpTime in its default datacollection-config.xml file.

> What would be the best way of replacing that OID for just windows
> machines?
>
> I am thinking that I have to pull the net-snmp include out of the
> Windows section, and then copy-paste the guts of the section back in
> with the OID that I want to change edited.
>
> Is there an easier way? (Or better yet, am I the only one seeing  
> strange
> uptime numbers on my Windows servers?)

I'm afraid you're making less sense the further down I read, which  
probably means that you started with a faulty assumption.  Perhaps you  
can provide actual examples of the values you're seeing for each of  
sysUpTime and hrSystemUptime, alongside the general kinds of values  
you're expecting?

-jeff

------------------------------------------------------------------------
-
This SF.Net email is sponsored by the Moblin Your Move Developer's
challenge
Build the coolest Linux based applications with Moblin SDK & win great
prizes
Grand prize is a trip for two to an Open Source event anywhere in the
world
http://moblin-contest.org/redirect.php?banner_id=100&url=/
_______________________________________________
Please read the OpenNMS Mailing List FAQ:
http://www.opennms.org/index.php/Mailing_List_FAQ

opennms-discuss mailing list

To *unsubscribe* or change your subscription options, see the bottom of
this page:
https://lists.sourceforge.net/lists/listinfo/opennms-discuss

-------------------------------------------------------------------------
This SF.Net email is sponsored by the Moblin Your Move Developer's challenge
Build the coolest Linux based applications with Moblin SDK & win great prizes
Grand prize is a trip for two to an Open Source event anywhere in the world
http://moblin-contest.org/redirect.php?banner_id=100&url=/
_______________________________________________
Please read the OpenNMS Mailing List FAQ:
http://www.opennms.org/index.php/Mailing_List_FAQ

opennms-discuss mailing list

To *unsubscribe* or change your subscription options, see the bottom of this page:
https://lists.sourceforge.net/lists/listinfo/opennms-discuss
Reply | Threaded
Open this post in threaded view
|

Re: Windows servers uptime values

(private) HKS
In reply to this post by Jeff Gehlbach
On Mon, Sep 8, 2008 at 2:52 PM, Jeff Gehlbach <[hidden email]> wrote:
<snip>
>> Be sure to remove mib2-host-resources-system from the Windows SNMP
>> system or you'll have two graphs, one of which is a tad confusing.
>
> Or set the "suppress" sub-property on the
> custom.windows.winSystemUptime report to "netsnmp.hrSystemUptime".


Hmm, I never noticed that sub-property. Handy. Thanks.

-HKS

-------------------------------------------------------------------------
This SF.Net email is sponsored by the Moblin Your Move Developer's challenge
Build the coolest Linux based applications with Moblin SDK & win great prizes
Grand prize is a trip for two to an Open Source event anywhere in the world
http://moblin-contest.org/redirect.php?banner_id=100&url=/
_______________________________________________
Please read the OpenNMS Mailing List FAQ:
http://www.opennms.org/index.php/Mailing_List_FAQ

opennms-discuss mailing list

To *unsubscribe* or change your subscription options, see the bottom of this page:
https://lists.sourceforge.net/lists/listinfo/opennms-discuss
Reply | Threaded
Open this post in threaded view
|

Re: Windows servers uptime values

James Zuelow
In reply to this post by (private) HKS
Thank you -- your custom graph and windows group is pretty similar to
what I had set up on my 1.2.9 install.

However even with the math to convert Microsoft's
millisecond/centisecond confusion, I think that you'll find that the
value wraps every 49 days 16 hours or so:

.1.3.6.1.2.1.1.3.0 (with old math): 59 Days, 21 Hours, 17 Minutes,
28 Seconds
.1.3.6.1.2.1.25.1.1 (with new math): 10 Days, 04 Hours, 14 Minutes,
58 Seconds

If you're monitoring a server that you know has been up longer than 50
days, please check that.  Since I can't count on my servers being
rebooted on patch Tuesdays, I'm still using  .1.3.6.1.2.1.1.3.0

:)


James Zuelow....................CBJ MIS (907)586-0236
Network Specialist...Registered Linux User No. 186591
-----Original Message-----
From: [hidden email]
[mailto:[hidden email]] On Behalf Of
(private) HKS
Sent: Monday, September 08, 2008 10:31 AM
To: General OpenNMS Discussion
Subject: Re: [opennms-discuss] Windows servers uptime values

1.3.6.1.2.1.25.1.1 is reported by Windows in milliseconds rather than
centisecond (as Net-SNMP reports it). I've gotten around this in a
mixed environment by removing mib2-host-resources-system from the
Windows system and replacing it with the windows-host group:

      <group name="windows-host" ifType="ignore">
        <mibObj oid=".1.3.6.1.2.1.25.3.3.1.2" instance="1"
alias="cpuPercentBusy" type="integer" />
        <!-- Memory in KBytes -->
        <mibObj oid=".1.3.6.1.2.1.25.2.2" instance="0"
alias="memorySize" type="integer" />
        <mibObj oid=".1.3.6.1.2.1.25.1.1" instance="0"
alias="winSystemUptime" type="timeticks" />
        <mibObj oid=".1.3.6.1.2.1.25.1.5" instance="0"
alias="hrSystemNumUsers" type="gauge" />
        <mibObj oid=".1.3.6.1.2.1.25.1.6" instance="0"
alias="hrSystemProcesses" type="gauge" />
      </group>

Then add the following graph to snmp-graph.properties:

report.custom.windows.winSystemUptime.name=System Uptime (Windows)
report.custom.windows.winSystemUptime.columns=winSystemUptime
report.custom.windows.winSystemUptime.type=nodeSnmp
report.custom.windows.winSystemUptime.command=--title="System Uptime" \
 --vertical-label Days \
 DEF:time={rrd1}:winSystemUptime:AVERAGE \
 CDEF:days=time,86400000,/ \
 LINE2:days#0000ff:"System Uptime (Days)" \
 GPRINT:days:AVERAGE:"Avg  \\: %8.1lf %s" \
 GPRINT:days:MIN:"Min  \\: %8.1lf %s" \
 GPRINT:days:MAX:"Max  \\: %8.1lf %s\\n"


Be sure to remove mib2-host-resources-system from the Windows SNMP
system or you'll have two graphs, one of which is a tad confusing.

-HKS


-------------------------------------------------------------------------
This SF.Net email is sponsored by the Moblin Your Move Developer's challenge
Build the coolest Linux based applications with Moblin SDK & win great prizes
Grand prize is a trip for two to an Open Source event anywhere in the world
http://moblin-contest.org/redirect.php?banner_id=100&url=/
_______________________________________________
Please read the OpenNMS Mailing List FAQ:
http://www.opennms.org/index.php/Mailing_List_FAQ

opennms-discuss mailing list

To *unsubscribe* or change your subscription options, see the bottom of this page:
https://lists.sourceforge.net/lists/listinfo/opennms-discuss
Reply | Threaded
Open this post in threaded view
|

Re: Windows servers uptime values

Mark Wolek-3
In reply to this post by James Zuelow
You could also use NSClient or SNMP-Informant to poll the perfmon
counter for Windows Uptime.

-----Original Message-----
From: [hidden email]
[mailto:[hidden email]] On Behalf Of
James Zuelow
Sent: Monday, September 08, 2008 2:07 PM
To: General OpenNMS Discussion
Subject: Re: [opennms-discuss] Windows servers uptime values

OK.  The values won't be identical as I'm doing them one after another.
But they should be within a few minutes.

One of my windows systems has been up for just about 60 days.  The
easiest way to get that into an e-mail is systeminfo output:

Microsoft Windows [Version 5.2.3790]
(C) Copyright 1985-2003 Microsoft Corp.

C:\Documents and Settings\james_zuelow>systeminfo

Host Name:                 CITY-DC1-W3S
OS Name:                   Microsoft(R) Windows(R) Server 2003, Standard
Edition

OS Version:                5.2.3790 Service Pack 2 Build 3790
OS Manufacturer:           Microsoft Corporation
OS Configuration:          Primary Domain Controller
OS Build Type:             Multiprocessor Free
Registered Owner:          CBJ
Registered Organization:   CBJ
Product ID:                69712-640-5317685-45192
Original Install Date:     4/27/2005, 3:40:50 PM
System Up Time:            59 Days, 20 Hours, 15 Minutes, 22 Seconds
System Manufacturer:       IBM

(snip)

Here are the two snmpwalk values:

DISMAN-EVENT-MIB::sysUpTimeInstance = Timeticks: (517059274) 59 days,
20:16:32.74

HOST-RESOURCES-MIB::hrSystemUptime.0 = Timeticks: (875770329) 101 days,
8:41:43.29

Now regardless of what the RFC's say, I **know** that this machine was
last rebooted about 2 months ago.  Not three and a half months ago.

I admit that I may be working from a faulty assumption, but I still want
to graph the first value as it is the accurate value for all of my
Windows machines.

I do not want to graph the second value, as it does not seem to have any
basis in reality for my Windows machines and gets further out of sync
with time.  It *is* accurate for my *nix machines however, so I only
want to replace the polled OID for Windows machines.  That is why I
wanted to edit datacollection-config.xml.

However I also want to stay as close to a stock install as possible, so
was wondering if there is an easier way.

James Zuelow....................CBJ MIS (907)586-0236
Network Specialist...Registered Linux User No. 186591

-----Original Message-----
From: [hidden email]
[mailto:[hidden email]] On Behalf Of Jeff
Gehlbach
Sent: Monday, September 08, 2008 10:42 AM
To: General OpenNMS Discussion
Subject: Re: [opennms-discuss] Windows servers uptime values

On Sep 8, 2008, at 2:19 PM, James Zuelow wrote:
> I seem to be getting crazy large values for uptime on my windows  
> servers
> with 1.5.93.
 >
> With my 1.2.9 installation I had edited datacollection-config.xml to
> poll .1.3.6.1.2.1.1.3.0 for uptime values.  Those values agreed with
> perfmon and systeminfo.

1.3.6.1.2.1.1.3 is sysUpTime.  It's defined in RFC1213 as:

  "The time (in hundredths of a second) since the network management  
portion of the system was last re-initialized."

That means that sysUpTime is the number of centiseconds since the SNMP  
agent on your Windows boxes was last started.  If you "net stop snmp ;  
net start snmp", then sysUpTime will go back to zero.  That means that  
sysUpTime is totally unreliable as an indicator of actual system uptime.

> The 1.5.93 installation is polling OID 1.3.6.1.2.1.25.1.1, which  
> results
> in values about (but not quite) twice as large.

1.3.6.1.2.1.25.1.1 is hrSystemUptime.  It's defined in RFC1514 (aka  
HOST-RESOURCES-MIB) as:

"The amount of time since this host was last initialized.  Note that  
this is different from sysUpTime in the SNMPv2-MIB [RFC1907] because  
sysUpTime is the uptime of the network management portion of the  
system."

The syntax given for hrSystemUptime is "TimeTicks", which means  
centiseconds, so the units are the same as for sysUpTime.

Any appearance of a correlation between the values of sysUpTime and  
hrSystemUptime is strictly coincidental, assuming the author of the  
SNMP agent from which you queried those values implemented both  
objects properly.

> The second OID is correct for Linux and AIX machines on our network,  
> but
> definitely not for Server 2003 or Server 2008.

Most modern SNMP agents support both sysUpTime and hrSystemUptime.

> With 1.2.9 it was easy to drop in the Windows uptime graph because  
> 1.2.9
> was not collecting any of the net-snmp OIDs for Windows servers.  So I
> could name the rrd the same as the net-snmp rrd, and the graph would
> magically appear.

Uhh, hrSystemUptime is not a "net-snmp OID".  Also, neither 1.2.9 nor  
1.5.93 includes sysUpTime in its default datacollection-config.xml file.

> What would be the best way of replacing that OID for just windows
> machines?
>
> I am thinking that I have to pull the net-snmp include out of the
> Windows section, and then copy-paste the guts of the section back in
> with the OID that I want to change edited.
>
> Is there an easier way? (Or better yet, am I the only one seeing  
> strange
> uptime numbers on my Windows servers?)

I'm afraid you're making less sense the further down I read, which  
probably means that you started with a faulty assumption.  Perhaps you  
can provide actual examples of the values you're seeing for each of  
sysUpTime and hrSystemUptime, alongside the general kinds of values  
you're expecting?

-jeff

------------------------------------------------------------------------
-
This SF.Net email is sponsored by the Moblin Your Move Developer's
challenge
Build the coolest Linux based applications with Moblin SDK & win great
prizes
Grand prize is a trip for two to an Open Source event anywhere in the
world
http://moblin-contest.org/redirect.php?banner_id=100&url=/
_______________________________________________
Please read the OpenNMS Mailing List FAQ:
http://www.opennms.org/index.php/Mailing_List_FAQ

opennms-discuss mailing list

To *unsubscribe* or change your subscription options, see the bottom of
this page:
https://lists.sourceforge.net/lists/listinfo/opennms-discuss

------------------------------------------------------------------------
-
This SF.Net email is sponsored by the Moblin Your Move Developer's
challenge
Build the coolest Linux based applications with Moblin SDK & win great
prizes
Grand prize is a trip for two to an Open Source event anywhere in the
world
http://moblin-contest.org/redirect.php?banner_id=100&url=/
_______________________________________________
Please read the OpenNMS Mailing List FAQ:
http://www.opennms.org/index.php/Mailing_List_FAQ

opennms-discuss mailing list

To *unsubscribe* or change your subscription options, see the bottom of
this page:
https://lists.sourceforge.net/lists/listinfo/opennms-discuss

-------------------------------------------------------------------------
This SF.Net email is sponsored by the Moblin Your Move Developer's challenge
Build the coolest Linux based applications with Moblin SDK & win great prizes
Grand prize is a trip for two to an Open Source event anywhere in the world
http://moblin-contest.org/redirect.php?banner_id=100&url=/
_______________________________________________
Please read the OpenNMS Mailing List FAQ:
http://www.opennms.org/index.php/Mailing_List_FAQ

opennms-discuss mailing list

To *unsubscribe* or change your subscription options, see the bottom of this page:
https://lists.sourceforge.net/lists/listinfo/opennms-discuss
Reply | Threaded
Open this post in threaded view
|

Re: Windows servers uptime values

(private) HKS
In reply to this post by James Zuelow
Hmmm, you're right, I didn't notice that. Thanks for pointing it out.

 I looked through the SNMP Informant MIB file and found the
system.systemUptime OID. It reports uptime in seconds, and appears to
be accurate.

datacollection-config.xml:
<mibObj oid=".1.3.6.1.4.1.9600.1.1.6" instance="0"
alias="winSystemUptime" type="guage" />

snmp-graph.properties:
report.custom.windows.winSystemUptime.name=System Uptime (Windows)
report.custom.windows.winSystemUptime.suppress=netsnmp.hrSystemUptime
report.custom.windows.winSystemUptime.columns=winSystemUptime
report.custom.windows.winSystemUptime.type=nodeSnmp
report.custom.windows.winSystemUptime.command=--title="System Uptime" \
 --vertical-label Days \
 DEF:time={rrd1}:winSystemUptime:AVERAGE \
 CDEF:days=time,86400,/ \
 LINE2:days#0000ff:"System Uptime (Days)" \
 GPRINT:days:AVERAGE:"Avg  \\: %8.1lf %s" \
 GPRINT:days:MIN:"Min  \\: %8.1lf %s" \
 GPRINT:days:MAX:"Max  \\: %8.1lf %s\\n"

-HKS


On Mon, Sep 8, 2008 at 4:19 PM, James Zuelow
<[hidden email]> wrote:

> Thank you -- your custom graph and windows group is pretty similar to
> what I had set up on my 1.2.9 install.
>
> However even with the math to convert Microsoft's
> millisecond/centisecond confusion, I think that you'll find that the
> value wraps every 49 days 16 hours or so:
>
> .1.3.6.1.2.1.1.3.0 (with old math):     59 Days, 21 Hours, 17 Minutes,
> 28 Seconds
> .1.3.6.1.2.1.25.1.1 (with new math):    10 Days, 04 Hours, 14 Minutes,
> 58 Seconds
>
> If you're monitoring a server that you know has been up longer than 50
> days, please check that.  Since I can't count on my servers being
> rebooted on patch Tuesdays, I'm still using  .1.3.6.1.2.1.1.3.0
>
> :)
>
>
> James Zuelow....................CBJ MIS (907)586-0236
> Network Specialist...Registered Linux User No. 186591
> -----Original Message-----
> From: [hidden email]
> [mailto:[hidden email]] On Behalf Of
> (private) HKS
> Sent: Monday, September 08, 2008 10:31 AM
> To: General OpenNMS Discussion
> Subject: Re: [opennms-discuss] Windows servers uptime values
>
> 1.3.6.1.2.1.25.1.1 is reported by Windows in milliseconds rather than
> centisecond (as Net-SNMP reports it). I've gotten around this in a
> mixed environment by removing mib2-host-resources-system from the
> Windows system and replacing it with the windows-host group:
>
>      <group name="windows-host" ifType="ignore">
>        <mibObj oid=".1.3.6.1.2.1.25.3.3.1.2" instance="1"
> alias="cpuPercentBusy" type="integer" />
>        <!-- Memory in KBytes -->
>        <mibObj oid=".1.3.6.1.2.1.25.2.2" instance="0"
> alias="memorySize" type="integer" />
>        <mibObj oid=".1.3.6.1.2.1.25.1.1" instance="0"
> alias="winSystemUptime" type="timeticks" />
>        <mibObj oid=".1.3.6.1.2.1.25.1.5" instance="0"
> alias="hrSystemNumUsers" type="gauge" />
>        <mibObj oid=".1.3.6.1.2.1.25.1.6" instance="0"
> alias="hrSystemProcesses" type="gauge" />
>      </group>
>
> Then add the following graph to snmp-graph.properties:
>
> report.custom.windows.winSystemUptime.name=System Uptime (Windows)
> report.custom.windows.winSystemUptime.columns=winSystemUptime
> report.custom.windows.winSystemUptime.type=nodeSnmp
> report.custom.windows.winSystemUptime.command=--title="System Uptime" \
>  --vertical-label Days \
>  DEF:time={rrd1}:winSystemUptime:AVERAGE \
>  CDEF:days=time,86400000,/ \
>  LINE2:days#0000ff:"System Uptime (Days)" \
>  GPRINT:days:AVERAGE:"Avg  \\: %8.1lf %s" \
>  GPRINT:days:MIN:"Min  \\: %8.1lf %s" \
>  GPRINT:days:MAX:"Max  \\: %8.1lf %s\\n"
>
>
> Be sure to remove mib2-host-resources-system from the Windows SNMP
> system or you'll have two graphs, one of which is a tad confusing.
>
> -HKS
>
>
> -------------------------------------------------------------------------
> This SF.Net email is sponsored by the Moblin Your Move Developer's challenge
> Build the coolest Linux based applications with Moblin SDK & win great prizes
> Grand prize is a trip for two to an Open Source event anywhere in the world
> http://moblin-contest.org/redirect.php?banner_id=100&url=/
> _______________________________________________
> Please read the OpenNMS Mailing List FAQ:
> http://www.opennms.org/index.php/Mailing_List_FAQ
>
> opennms-discuss mailing list
>
> To *unsubscribe* or change your subscription options, see the bottom of this page:
> https://lists.sourceforge.net/lists/listinfo/opennms-discuss
>

-------------------------------------------------------------------------
This SF.Net email is sponsored by the Moblin Your Move Developer's challenge
Build the coolest Linux based applications with Moblin SDK & win great prizes
Grand prize is a trip for two to an Open Source event anywhere in the world
http://moblin-contest.org/redirect.php?banner_id=100&url=/
_______________________________________________
Please read the OpenNMS Mailing List FAQ:
http://www.opennms.org/index.php/Mailing_List_FAQ

opennms-discuss mailing list

To *unsubscribe* or change your subscription options, see the bottom of this page:
https://lists.sourceforge.net/lists/listinfo/opennms-discuss
Reply | Threaded
Open this post in threaded view
|

Re: Windows servers uptime values

Jeff Gehlbach
In reply to this post by James Zuelow
On Sep 8, 2008, at 4:19 PM, James Zuelow wrote:

> However even with the math to convert Microsoft's
> millisecond/centisecond confusion, I think that you'll find that the
> value wraps every 49 days 16 hours or so:

Note that sysUpTime is not immune to counter wraps either, you'll see  
one every 497 days or so on every platform.

-jeff

-------------------------------------------------------------------------
This SF.Net email is sponsored by the Moblin Your Move Developer's challenge
Build the coolest Linux based applications with Moblin SDK & win great prizes
Grand prize is a trip for two to an Open Source event anywhere in the world
http://moblin-contest.org/redirect.php?banner_id=100&url=/
_______________________________________________
Please read the OpenNMS Mailing List FAQ:
http://www.opennms.org/index.php/Mailing_List_FAQ

opennms-discuss mailing list

To *unsubscribe* or change your subscription options, see the bottom of this page:
https://lists.sourceforge.net/lists/listinfo/opennms-discuss
Reply | Threaded
Open this post in threaded view
|

Re: Windows servers uptime values

(private) HKS
In reply to this post by (private) HKS
Whoops, guage should read "gauge"

-HKS

On Mon, Sep 8, 2008 at 4:34 PM, (private) HKS <[hidden email]> wrote:

> Hmmm, you're right, I didn't notice that. Thanks for pointing it out.
>
>  I looked through the SNMP Informant MIB file and found the
> system.systemUptime OID. It reports uptime in seconds, and appears to
> be accurate.
>
> datacollection-config.xml:
> <mibObj oid=".1.3.6.1.4.1.9600.1.1.6" instance="0"
> alias="winSystemUptime" type="guage" />
>
> snmp-graph.properties:
> report.custom.windows.winSystemUptime.name=System Uptime (Windows)
> report.custom.windows.winSystemUptime.suppress=netsnmp.hrSystemUptime
> report.custom.windows.winSystemUptime.columns=winSystemUptime
> report.custom.windows.winSystemUptime.type=nodeSnmp
> report.custom.windows.winSystemUptime.command=--title="System Uptime" \
>  --vertical-label Days \
>  DEF:time={rrd1}:winSystemUptime:AVERAGE \
>  CDEF:days=time,86400,/ \
>  LINE2:days#0000ff:"System Uptime (Days)" \
>  GPRINT:days:AVERAGE:"Avg  \\: %8.1lf %s" \
>  GPRINT:days:MIN:"Min  \\: %8.1lf %s" \
>  GPRINT:days:MAX:"Max  \\: %8.1lf %s\\n"
>
> -HKS
>
>
> On Mon, Sep 8, 2008 at 4:19 PM, James Zuelow
> <[hidden email]> wrote:
>> Thank you -- your custom graph and windows group is pretty similar to
>> what I had set up on my 1.2.9 install.
>>
>> However even with the math to convert Microsoft's
>> millisecond/centisecond confusion, I think that you'll find that the
>> value wraps every 49 days 16 hours or so:
>>
>> .1.3.6.1.2.1.1.3.0 (with old math):     59 Days, 21 Hours, 17 Minutes,
>> 28 Seconds
>> .1.3.6.1.2.1.25.1.1 (with new math):    10 Days, 04 Hours, 14 Minutes,
>> 58 Seconds
>>
>> If you're monitoring a server that you know has been up longer than 50
>> days, please check that.  Since I can't count on my servers being
>> rebooted on patch Tuesdays, I'm still using  .1.3.6.1.2.1.1.3.0
>>
>> :)
>>
>>
>> James Zuelow....................CBJ MIS (907)586-0236
>> Network Specialist...Registered Linux User No. 186591
>> -----Original Message-----
>> From: [hidden email]
>> [mailto:[hidden email]] On Behalf Of
>> (private) HKS
>> Sent: Monday, September 08, 2008 10:31 AM
>> To: General OpenNMS Discussion
>> Subject: Re: [opennms-discuss] Windows servers uptime values
>>
>> 1.3.6.1.2.1.25.1.1 is reported by Windows in milliseconds rather than
>> centisecond (as Net-SNMP reports it). I've gotten around this in a
>> mixed environment by removing mib2-host-resources-system from the
>> Windows system and replacing it with the windows-host group:
>>
>>      <group name="windows-host" ifType="ignore">
>>        <mibObj oid=".1.3.6.1.2.1.25.3.3.1.2" instance="1"
>> alias="cpuPercentBusy" type="integer" />
>>        <!-- Memory in KBytes -->
>>        <mibObj oid=".1.3.6.1.2.1.25.2.2" instance="0"
>> alias="memorySize" type="integer" />
>>        <mibObj oid=".1.3.6.1.2.1.25.1.1" instance="0"
>> alias="winSystemUptime" type="timeticks" />
>>        <mibObj oid=".1.3.6.1.2.1.25.1.5" instance="0"
>> alias="hrSystemNumUsers" type="gauge" />
>>        <mibObj oid=".1.3.6.1.2.1.25.1.6" instance="0"
>> alias="hrSystemProcesses" type="gauge" />
>>      </group>
>>
>> Then add the following graph to snmp-graph.properties:
>>
>> report.custom.windows.winSystemUptime.name=System Uptime (Windows)
>> report.custom.windows.winSystemUptime.columns=winSystemUptime
>> report.custom.windows.winSystemUptime.type=nodeSnmp
>> report.custom.windows.winSystemUptime.command=--title="System Uptime" \
>>  --vertical-label Days \
>>  DEF:time={rrd1}:winSystemUptime:AVERAGE \
>>  CDEF:days=time,86400000,/ \
>>  LINE2:days#0000ff:"System Uptime (Days)" \
>>  GPRINT:days:AVERAGE:"Avg  \\: %8.1lf %s" \
>>  GPRINT:days:MIN:"Min  \\: %8.1lf %s" \
>>  GPRINT:days:MAX:"Max  \\: %8.1lf %s\\n"
>>
>>
>> Be sure to remove mib2-host-resources-system from the Windows SNMP
>> system or you'll have two graphs, one of which is a tad confusing.
>>
>> -HKS
>>
>>
>> -------------------------------------------------------------------------
>> This SF.Net email is sponsored by the Moblin Your Move Developer's challenge
>> Build the coolest Linux based applications with Moblin SDK & win great prizes
>> Grand prize is a trip for two to an Open Source event anywhere in the world
>> http://moblin-contest.org/redirect.php?banner_id=100&url=/
>> _______________________________________________
>> Please read the OpenNMS Mailing List FAQ:
>> http://www.opennms.org/index.php/Mailing_List_FAQ
>>
>> opennms-discuss mailing list
>>
>> To *unsubscribe* or change your subscription options, see the bottom of this page:
>> https://lists.sourceforge.net/lists/listinfo/opennms-discuss
>>
>

-------------------------------------------------------------------------
This SF.Net email is sponsored by the Moblin Your Move Developer's challenge
Build the coolest Linux based applications with Moblin SDK & win great prizes
Grand prize is a trip for two to an Open Source event anywhere in the world
http://moblin-contest.org/redirect.php?banner_id=100&url=/
_______________________________________________
Please read the OpenNMS Mailing List FAQ:
http://www.opennms.org/index.php/Mailing_List_FAQ

opennms-discuss mailing list

To *unsubscribe* or change your subscription options, see the bottom of this page:
https://lists.sourceforge.net/lists/listinfo/opennms-discuss
Reply | Threaded
Open this post in threaded view
|

Re: Windows servers uptime values

Jeff Gehlbach
On Sep 8, 2008, at 4:49 PM, (private) HKS wrote:

> Whoops, guage should read "gauge"

Actually it should read "counter", because despite the MIB definition  
erroneously declaring it as a Gauge32, it's a monotonically  
increasing  value.  The various INFORMANT-* MIBs tend to be full of  
this particular error.  Collecting it as a counter will give the  
expected behavior in openNMS.

-jeff

>
>
> -HKS
>
> On Mon, Sep 8, 2008 at 4:34 PM, (private) HKS  
> <[hidden email]> wrote:
>> Hmmm, you're right, I didn't notice that. Thanks for pointing it out.
>>
>> I looked through the SNMP Informant MIB file and found the
>> system.systemUptime OID. It reports uptime in seconds, and appears to
>> be accurate.
>>
>> datacollection-config.xml:
>> <mibObj oid=".1.3.6.1.4.1.9600.1.1.6" instance="0"
>> alias="winSystemUptime" type="guage" />

-------------------------------------------------------------------------
This SF.Net email is sponsored by the Moblin Your Move Developer's challenge
Build the coolest Linux based applications with Moblin SDK & win great prizes
Grand prize is a trip for two to an Open Source event anywhere in the world
http://moblin-contest.org/redirect.php?banner_id=100&url=/
_______________________________________________
Please read the OpenNMS Mailing List FAQ:
http://www.opennms.org/index.php/Mailing_List_FAQ

opennms-discuss mailing list

To *unsubscribe* or change your subscription options, see the bottom of this page:
https://lists.sourceforge.net/lists/listinfo/opennms-discuss
Reply | Threaded
Open this post in threaded view
|

Re: Windows servers uptime values

James Zuelow
In reply to this post by Jeff Gehlbach
Luckily I don't have any servers that have been up THAT long.

I think I can ensure a patch cycle at least every 365 days, even on the
*nix boxes.

This is primarily why I'm interested in the Windows uptime to begin
with, and why I was interested in high thresholds and flagging a machine
that has an uptime of greater than 60 or 90 days.  That would tell me at
a glance that a server was out of date.

If the SNMP informant value is in seconds, then it would be decades
before it wrapped, right?

James Zuelow....................CBJ MIS (907)586-0236
Network Specialist...Registered Linux User No. 186591
-----Original Message-----
From: [hidden email]
[mailto:[hidden email]] On Behalf Of Jeff
Gehlbach
Sent: Monday, September 08, 2008 12:46 PM
To: General OpenNMS Discussion
Subject: Re: [opennms-discuss] Windows servers uptime values

On Sep 8, 2008, at 4:19 PM, James Zuelow wrote:

> However even with the math to convert Microsoft's
> millisecond/centisecond confusion, I think that you'll find that the
> value wraps every 49 days 16 hours or so:

Note that sysUpTime is not immune to counter wraps either, you'll see  
one every 497 days or so on every platform.

-jeff

------------------------------------------------------------------------
-
This SF.Net email is sponsored by the Moblin Your Move Developer's
challenge
Build the coolest Linux based applications with Moblin SDK & win great
prizes
Grand prize is a trip for two to an Open Source event anywhere in the
world
http://moblin-contest.org/redirect.php?banner_id=100&url=/
_______________________________________________
Please read the OpenNMS Mailing List FAQ:
http://www.opennms.org/index.php/Mailing_List_FAQ

opennms-discuss mailing list

To *unsubscribe* or change your subscription options, see the bottom of
this page:
https://lists.sourceforge.net/lists/listinfo/opennms-discuss

-------------------------------------------------------------------------
This SF.Net email is sponsored by the Moblin Your Move Developer's challenge
Build the coolest Linux based applications with Moblin SDK & win great prizes
Grand prize is a trip for two to an Open Source event anywhere in the world
http://moblin-contest.org/redirect.php?banner_id=100&url=/
_______________________________________________
Please read the OpenNMS Mailing List FAQ:
http://www.opennms.org/index.php/Mailing_List_FAQ

opennms-discuss mailing list

To *unsubscribe* or change your subscription options, see the bottom of this page:
https://lists.sourceforge.net/lists/listinfo/opennms-discuss