Monitor HP insight ?

[Kai-Kai]

Hello,

I'm currently testing Zabbix on a network with Unix and Windows servers. On the Unix servers, I've installed the zabbix_agent but on the Windows is already installed and used "HP insight".

So, I'm looking for a way to make Zabbix and Insight work together. Before trying to do my own scripts, I would to like to know if anything already exists around this topic ?

I've found http://www.zabbix.com/wiki/doku.php?...ates&s=insight where a template seems to be for insight (I don't currently know how it works).
Has anybody anything to say around my search ?

Thanks in advance.

[Cray]

Hi !

I'm currently mainly working with HP Proliant servers, so I've been using the HP insights template actively.

By the way, the template you're mentionning is already included in Zabbix 1.6.x

This is how it works (on a windows server): if you are planning to get the HP insight manager informations trough Zabbix (hardware status etc), you have to install the HP proliant support pack (latest version) and configure the SNMP service to read all the additionnal CPQHLTH MIB.

Once it's done, you can link the server with the Insight manager template, and you'll get the hardware monitoring of your server

[Kai-Kai]

Thanks a lot for your answer.

I think my zabbix is not currently configured properly to receive snmp traps and the documentation isn't really clear concerning this topic...

Could you explain me how to configure zabbix well to receive snmp traps so that I can try to receive the hp insight one ?
In fact, I'm not sure my server, running under solaris 10, receives snmp traps...

[Cray]

You don't need the SNMP traps to get the informations from HP insight (I think you misunderstand how HP insight agents work).

Here's a brief explanation on zabbix agent and HP insight :

- the zabbix_agentd helps collecting data for the Zabbix server and can be queried directly by the Zabbix server

- On the other hand, the HP insight agents collect data and make it available trough SNMP MIBs (The CPQHLTH MIB is one of the many HP MIB that can be queried)

So, you can use the informations given by the zabbix_agentd AND the informations given by the HP insight trough SNMP MIBs (the informations collected by the HP insight agents can give you informations on your server health, and the zabbix_agentd can give you information on your OS etc).

To get the informations collected by the HP Insight agents, just make sure the SNMP services are running and properly configured with the community name you want, and then link your host with the HP_Insight template in zabbix.



If you wanna go deeper, you can use SNMP traps with the HP iLO (but its another story )

[Kai-Kai]

Perhaps I had misunderstood yes. Thanks a lot for this explaination !

In fact what I would like to do can be devided into 3 parts :

* the first one is : I would like to obtain the summary of insight, on each machine. Simply : "OK" or "Problem". If the status is ok, ok... if the status is problem : alert on dashboard.
I absolutely need that.

The 2 other parts are optionnal, but could be great :
* the second one : it could be interesting to be able to query hp insight agent, using snmp or anything else, to obtain infos about several things on the servers.
* the last one : it could be very interesting if my zabbix could collect snmp datas from insight or other equipments to generate alerts.

I know the bases of snmp : how it works (mibs, traps, queries)... but not really how to use / configure it on solaris 10, and in zabbix 1.6...

Do you think you can help me ?

I go back trying to make snmp work on my solaris server.

First, do you know if I should use snmpd or snmptrapd ?

Thanks in advance for your help.

[Cray]

* the first one : [...]
* the second one : [...]
* the last one : [...]

All the points you mention will lead to the same solution : all you have to do is to query the HP insight agents trough SNMP.

First, do you know if I should use snmpd or snmptrapd ?

You don't have to use those SNMP daemons to work out what you wan't to do (these daemons are for SNMP traping only as far as I remember).


- is the snmp package installed on your platform ? (Solaris is a bit different than Linux, I know, but an snmp package should be available anyway)
- Have you compiled Zabbix with the "--with-net-snmp" option ? (you can't compile it that way if the snmp package is not installed)

That's it, if you have compiled Zabbix along with the SNMP support, you should be able to query the HP insight agents right away

[Kai-Kai]

Thanks for your answer.

All the points you mention will lead to the same solution : all you have to do is to query the HP insight agents trough SNMP.

I think there's a small difference between querying the insight machine and waiting for traps from it.

I think you're right : first one and third one are the same.
So let's try to do that.

is the snmp package installed on your platform ?

Yes. I've installed the package from sunfreeware.com :
ftp://ftp.sunfreeware.com/pub/freewa...sparc-local.gz

- Have you compiled Zabbix with the "--with-net-snmp" option ?

I've compiled the server with --with-net-snmp=/usr/local/bin/net-snmp-config.

In the web interface, I've used "insight template" for my windows machine and some agent based tests. The agent based tests run ok, but as far as the snmp infos are concerned : nothing.
In the screen "dernières métriques" (something like "last values" in english I think), the values are empty...

Is there a simple way to test if snmp work in zabbix ? To know if the problem come from zabbix, the snmp package, other thing... ?

Thanks in advance.

[Kai-Kai]

Oh oh !
It seems to work... ^^

I decided to search the problem away... so I've verified the configuration of the insight itself... and the man who installed the hp insight has changed the name of the snmp comunity...

I've changed it in zabbix too... and I've now some values which are coming to my zabbix. Wonderfull !

I'll try to right understand the values I can monitore this way and if all is ok, I'll come to say here all is ok for me.

Thanks again for your help.

[Kai-Kai]

The values currently collected are "1" and "2". According to the trigger, it's ok, as the trigger make an alert if value > 2...

Do you know more preciesly what these values mean ?

[Cray]

You're welcome

Regarding the use of SNMP traps, as mentionned before, you'll have to use the HP iLO/2 interface (at least it's what I'm doing with the proliant servers) :

You can configure the iLO/2 to send traps to the Zabbix server (or proxy) whenever the server restarts or a special 'low-level' event occurs.

Might be useful, but in this case, you'll have to setup SNMP trapping for zabbix on the server / proxy side.

[Kai-Kai]

Do you know where I can find more informations about the meaning of the different status return by the snmp queries ?

For example, the insight template in zabbix defines several tests, returning 1 or 2 to my server... and there are also some triggers, if value is more than 2.

But what it more preciesly the meaning of these codes ? What can be all the codes ?

I think it's usefull to understand that to make my own triggers.

Thanks.

[Cray]

I did some researches regarding the CPQHLTH MIB (which I was mostly interested in because it contains al the server health infos).

You can find this MIB at this point of the SNMP tree :

iso.org.dod.internet.private.enterprises.compaq.cp qHealth

Depending on the tool you're using to do the SNMP walk / browsing, you'll have some info concerning the codes given by HP (e.g : 2 == ok, 3 == problem, 4 == failure etc)

[Kai-Kai]

I'm sorry for this question which is I think a bit stupide, but I'm completly newbie in the domain of SNMP... : could you say me how I can browse the MIB tree to find the meaning of the values ?
Is there a soft to do that ? A website ?

[Kai-Kai]

Hello.
I come back here with my question... I'haven't find any answer...

When I query the serveur, I receive a return code... but still no idea of the meaning of these codes.

Example, as far as the "global status" is concerned :

Command :
snmpget -v 2c -c <COMUNITY> -r 2 -t 20 -m ALL -O T <SERVER_IP> .1.3.6.1.4.1.232.6.1.3.0

Answer :
SNMPv2-SMI::enterprises.232.6.1.3.0 = INTEGER: 2

The oid : .1.3.6.1.4.1.232.6.1.3.0 is the one configured in Zabbix for "global status" check.

Can anybody help me ?

[Cray]

Here you go :

http://www.kilala.nl/Sysadmin/index.php?id=736 (have a look at the 'value' column)

http://www.kilala.nl/Sysadmin/Images/HPMIB.pdf

These 2 links should provide you the information you are looking for about the code status