gjhvjv Absent Member.
Absent Member.
376 views

Problem in NMM

Jump to solution
Hi, my name is Marc and i'm currently doing a large project using Network Node Manager. i managed to install NNM and finish my dicovery of the network diagram.
I have a few questions firstly, after mine discovery of the network, i begin to drill down to one of the subnets and managed to see quite a number of nodes. Under the nodes, there are names under it but what does the name stands for? Is it the computer name? Something strange is also happening like some nodes will suddenly disappear. What does it means?
And lastly, after i done the dicovery, what will i have to do next in order to make full use of NNM to manage and monitor the network. Thankz.
newbie
0 Likes
1 Solution

Accepted Solutions
Marcel Kuiper Absent Member.
Absent Member.

Re: Problem in NMM

Jump to solution
A device needs at least 1 ip address or else it would be impossible to talk SNMP (netmon decides on up/down status by pinging or SNMP status)

Empty submaps can have several reasons

1 - The addresses that live in the subnet don't pass $OV_CONF/netmon.noDiscover or don't pass the discovery filter. Filters are all defined in $OV_CONF/C/filters; the discovery filter is defined under OPTIONS-->Network Polling Configuration on the GENERAL TAB; if there is no filter name and/or the radio button is not pressed, you don't use a discovery filter

2 - The devices have not yet been discovered. NNM does not search the ip address space linear. It depends on hints. It will get hints from ARP caches of devices it manages and route (next hop) entries. If the devices are just not in the ARP cache of the router when netmon is doing a SNMP poll then netmon has no hints. You can use the loadhosts command to enter devices manually from the commandline

3 - There's a firewall/packetfilter (or router ACL) that blocks ICMP and/or SNMP traffic to/from the subnet


ABout configuring traps. First of all you must configure your devices to send traps to the NNM machine. Second you must load the appropriate MIB's for the devices and import the TRAP definitions that are in the MIB's into the $OV_CONF/C/trapd.conf (use OPTIONS-->Load/Unload MIB's menu). Last you must configure the events so that it displays in the alarm browser at least (use OPTIONS-->Event Configuration menu)

About perl. NNM installs perl 5.6.0 with NNM 6.2 You will find the executable as /opt/OV/bin/Perl/bin/perl

Maybe for writing your own scripts you want to compile/install the latest stable release from www.CPAN.org under /usr/local. You will find a lot of useful libraries there as well

Marcel
If we can't fix it, it ain't broken
21 Replies
Marcel Kuiper Absent Member.
Absent Member.

Re: Problem in NMM

Jump to solution
First of all I advise you to read the "maage your network" guide. Get the pdf from openview.hp.com

The namegiving follows a special algorithm
From the same site as above you can get the "netmon internals" document under development resources. From the top of my head:

- if software loopback addresses can be resolved the name is taken from the lowest numbered software loopback ip

- else the name is taken from the lowest numbered ip that can be resolved

- else SNMP system name is used

- else lowest numbered software loopback ip

etc

Devices will dissappear or will not get on the mapfor a few reasons:

- According to oid_to_type they belong on an other submap then the Internet submap (i.e. netowk layer or segments layer or node layer)

- They could not be polled for a while (default remove after a week)

- A discovery filter or netmon.noDiscover line tells not to manage that device or ip

Next thing to do:

- Perhaps you want to adjust your map layout by adding containers on the Internet map and dragging and dropping devices to the right location

- Configure relevant events for your environment (OV_Node_Down, OV_IF_Down, SNMP_Link_Down, device specific traps)

That's enough work for a while

Cheers

MArcel
If we can't fix it, it ain't broken
Highlighted
Alan Deger Absent Member.
Absent Member.

Re: Problem in NMM

Jump to solution
The names should be the computer or node names. They should be picked up from a DNS query of the IP address of the node or the SNMP system name. Maybe someone else can address the "disappearing nodes" problem...

Your next problem isn't with NNM--it's with the specifications of your "large project". You have a lot of homework to do that maybe should have been done (or that you haven't mentioned doing) before you even uncorked NNM. I might suggest as a first step examining your network from what you know it to be (not necessarily from how NNM has discovered it) and start to prioritize nodes into three groups of highest (must be up 24x7) medium (must be up during business hours) and low (can be down during business hours). Divide these into monitoring classes (usually based on what the thing does or is: computer, switch, router, etc) and start reading some "best practices" papers on how to monitor and manage the things. As an example of this, we're a Cisco shop for routing and switching so it's easy to find best practices white papers on the Cisco site which will always have suggestions for how and what to manage with an NMS (network management system of which NNM is just one). Do the same for your computers and try to devise health checks for your major services (usually scripts that read a web page, send and pick up email, etc).
Consider who needs to see what of these nodes in NNM and start building NNM maps of these (but keep them to a minimum).

Give some consideration as to how you're going to alert on problems you've decided to monitor for (Perl scripts, shell scripts, Telalert, etc. via pager, cell phone, 2-way text messaging, etc.)

After all of this then start reading the manuals and (as your next resort) posting questions to the forum that are specific enough for us to get our teeth into (e.g., "How do I baseline and set thresholds for CPU utilization on my xxx router")

Usually posing a well thought out question is half the battle to answering it yourself or gettting someone else to answering it.

Hope this helps,
ard
Seun Ewulomi_1 Absent Member.
Absent Member.

Re: Problem in NMM

Jump to solution
Hi,

The guys have giving very good advise. You need to plan. I always say this whenever I get the oppurtunity to say so "KNOW YOUR ENVIRONMENT".
Auto discovery is very good but becomes useless if you dont know what you expect to be discovered, it just leads to more confusion.

People seem to defeat how to properly use autodiscovery Im afraid to say personally that HP should properly define how autodiscovery should be properly used. The manuals just literally imply that use autodiscovery to discover your network and NNM willl start showing you the gaps in your well configured network scheme. To know that NNM is wrong in some layouts during discovery(or to know the gaps as implied)you need to know the network

AFAIK you can get away with staright autodiscovery with no planning in small-medium environment but in large environments you need to plan carefully suggestions that come to mind:

filters (e.g discovery filters)
oid_to_type file
netmon.Nodiscover file
oid_to_sym file
Dns

and various other things as well.

Anyway enough of that.

The past thread below might help in which we discussed similar/related issues

Installation and implementation
http://forums.itrc.hp.com/cm/QuestionAnswer/1,,0x4e0d543254bfd611abdb0090277a778c,00.html

regards,
gab
Jesus Christ is LORD
0 Likes
gjhvjv Absent Member.
Absent Member.

Re: Problem in NMM

Jump to solution
Hi you guys are very helpful indeed. Thankz for all your help.
I need to ask u all some other questions.
Firstly, when i view the nodes many are in dirrerent colours. I know what the colour stand for. For example when nodes are in warning minor or even major status, they tend to contain multiple IP addresses in a single node. Does that means that each node should at most contain one ip address?
Second question is after i discover the network, i began drilling down to the submaps. i noticed that there are like 11 subnets connected using 4 gateways and the lines connecting to the subnets all show different VLAN addresses. I realised that i can only go into one of the subnets and see the nodes inside that subnet. whereas other subnets when i double clicked it shows an empty submap. Why is this so? Is it because i can only managed to view one of the subnets.
Lastly, how do you configure events like setting devices traps on switch or routers.
And for doing scripting, like for example pearl script, where can u write it? Is it inside the NNM software?
Thankz.
newbie
0 Likes
Marcel Kuiper Absent Member.
Absent Member.

Re: Problem in NMM

Jump to solution
A device needs at least 1 ip address or else it would be impossible to talk SNMP (netmon decides on up/down status by pinging or SNMP status)

Empty submaps can have several reasons

1 - The addresses that live in the subnet don't pass $OV_CONF/netmon.noDiscover or don't pass the discovery filter. Filters are all defined in $OV_CONF/C/filters; the discovery filter is defined under OPTIONS-->Network Polling Configuration on the GENERAL TAB; if there is no filter name and/or the radio button is not pressed, you don't use a discovery filter

2 - The devices have not yet been discovered. NNM does not search the ip address space linear. It depends on hints. It will get hints from ARP caches of devices it manages and route (next hop) entries. If the devices are just not in the ARP cache of the router when netmon is doing a SNMP poll then netmon has no hints. You can use the loadhosts command to enter devices manually from the commandline

3 - There's a firewall/packetfilter (or router ACL) that blocks ICMP and/or SNMP traffic to/from the subnet


ABout configuring traps. First of all you must configure your devices to send traps to the NNM machine. Second you must load the appropriate MIB's for the devices and import the TRAP definitions that are in the MIB's into the $OV_CONF/C/trapd.conf (use OPTIONS-->Load/Unload MIB's menu). Last you must configure the events so that it displays in the alarm browser at least (use OPTIONS-->Event Configuration menu)

About perl. NNM installs perl 5.6.0 with NNM 6.2 You will find the executable as /opt/OV/bin/Perl/bin/perl

Maybe for writing your own scripts you want to compile/install the latest stable release from www.CPAN.org under /usr/local. You will find a lot of useful libraries there as well

Marcel
If we can't fix it, it ain't broken
gjhvjv Absent Member.
Absent Member.

Re: Problem in NMM

Jump to solution
Thanks for your advice.
Regarding mine first problem, i noticed on mine NNM station that one of the computers beside me shows minor status and when i double click on that node, it shows 11 ip addresses inside that node. So i type ipconfig in command prompt on that computer and it shows me only 1 ip address which is on the computer. Why is his so?
Really appreaciate yr effort.
🙂
newbie
0 Likes
Marcel Kuiper Absent Member.
Absent Member.

Re: Problem in NMM

Jump to solution
It is possible that all the ip's resolve to the same name or that the SNMP sysname on all these ip addresses is the same (stupid I know because NNM could check easily by comparing against the ipAddrTable)

It is also possible that these addresses were once on the Node and have yet not been deleted

Last but not least are these all differen addresses or all the same, because it is possbile that you are looking at copies (are a lot of th interfaces with a shadow arounfd it?)

BTW, You are allowed to rate the appreciation 😉
If we can't fix it, it ain't broken
0 Likes
gjhvjv Absent Member.
Absent Member.

Re: Problem in NMM

Jump to solution
ok. I will rate them. 🙂
erm regarding yr question, the ip addresses are all different. There are 11 ip addresses but only one is that computer address.
Also regarding the empty sunmaps. Lets say that theres is no firewall that blocks the incoming traffic. How can configure NNM to let it shows all the nodes on every subnets.
newbie
0 Likes
gjhvjv Absent Member.
Absent Member.

Re: Problem in NMM

Jump to solution
I have another question regarding the alarm browser. It shows me an error which is: Your name services are performing poorly, lkely misconfigured. The average for the past 101 lookups took 4066.7 milliseconds, expect less than 500 mlliseconds.
What does this error means?
Thanksz.
newbie
0 Likes
Marcel Kuiper Absent Member.
Absent Member.

Re: Problem in NMM

Jump to solution
About the 11 addresses

$OV_BIN/snmpget address .1.3.6.1.2.1.1.5.0

(SNMP systemname) are they all the same?

ABout DNS. Your DNS is probably misconfigured. NNM is experiencing a lot of timeouts on DNS lookup's (can be around 75 secs per lookup) which impacts netmon's performance badly

Use the checkDNS.ovpl utility to search for objects with ip-addresses that have a large lookup time (larger then 1 second is already suspect on a LAN)

Marcel
If we can't fix it, it ain't broken
gjhvjv Absent Member.
Absent Member.

Re: Problem in NMM

Jump to solution
okok. i will check that.
I install NNM on another system and when i tried to open it, it shows me two errors messages:
Fatal IPMap Error-unable to connect to ovtopmd
Reason- Cannot connect to it786p38(address=152.226.238.155, port=2532): connection refused.
Why is this so? and how do it solve it? Need it urgent. Thanksz.
newbie
0 Likes
gjhvjv Absent Member.
Absent Member.

Re: Problem in NMM

Jump to solution
okok. i will check that.
I install NNM on another system and when i tried to open it, it shows me two errors messages:
Fatal IPMap Error-unable to connect to ovtopmd
Reason- Cannot connect to it786p38(address=152.226.238.155, port=2532): connection refused.
Why is this so? and how do it solve it? Need it urgent. Thanksz.
newbie
0 Likes
gjhvjv Absent Member.
Absent Member.

Re: Problem in NMM

Jump to solution
I have another Question which is inside the NNM there is one function which is under performance/system/disk space
Is this function for checking the disk space of other computers? i tried it to check the computers beside me but it shows me error that no response arrived before timeout.
Why is this so? and how can i solve it?
Thanksz a lot.
newbie
0 Likes
Marcel Kuiper Absent Member.
Absent Member.

Re: Problem in NMM

Jump to solution
The hosts need to run an agent that implements the HOST-RESOURCES MIB

Marcel
If we can't fix it, it ain't broken
0 Likes
The opinions expressed above are the personal opinions of the authors, not of Micro Focus. By using this site, you accept the Terms of Use and Rules of Participation. Certain versions of content ("Material") accessible here may contain branding from Hewlett-Packard Company (now HP Inc.) and Hewlett Packard Enterprise Company. As of September 1, 2017, the Material is now offered by Micro Focus, a separately owned and operated company. Any reference to the HP and Hewlett Packard Enterprise/HPE marks is historical in nature, and the HP and Hewlett Packard Enterprise/HPE marks are the property of their respective owners.