SNMP in Zabbix 2.0

[Ishan]

Hello Aib

I have gone through this document before to create the discovery rule, May be i am doing something wrong

Which SNMP ODI can use to create the discovery rule in my case? i am bit confused on that.

Thanks & Regards,
Ishan

[aib]

I have gone through this document before to create the discovery rule, May be i am doing something wrong

Which SNMP ODI can use to create the discovery rule in my case? i am bit confused on that.

I did translation from text form of OID to numeric to be more clear.

# snmptranslate -On SNMPv2-SMI::enterprises.5003.10.10.2.21.1.3.0
.1.3.6.1.4.1.5003.10.10.2.21.1.3.0

Then you can see, that part 1.3.6.1.4.1.5003.10.10.2.21.1.3.0 is the constant and only last digit changes (0,1,2) when you do snmpwalk.

We create the Discovery Rule where
SNMP OID: 1.3.6.1.4.1.5003.10.10.2.21.1.3.0

In this Discovery Rule we create one Item Prototype:
Name: Channel $1 load
Type: SNMPv2 agent
Key: channel.load["{#SNMPVALUE}"]
Host interface: {use correct host IP and port}
SNMP OID: 1.3.6.1.4.1.5003.10.10.2.21.1.3.0.{#SNMPINDEX}
SNMP community: {use your community OR configure macros {#SNMP_COMMUNITY}}


After your Discovery rule works you will have three Items with keys:
- channel.load["0"]
- channel.load["1"]
- channel.load["2"]

And then you can manually create a graph and add all that key into one graph.

I'm not sure but it looks like you have two {devices/trunks} and each have three {connections/interfaces}
But when snmpwalk scanned that interfaces, only first device shows some load. Because of that I told you to create a rule only for first device.
If you like, you can create one more Discovery rule and use SNMP OID 1.3.6.1.4.1.5003.10.10.2.21.1.3.1
And when you will create Item prototype, use the same , a little bit different OID for the second channel.

I don't want to explain you something about "Dynamic SNMP" because I read some complains that is not working properly yet. (High load for Zabbix Server CPU, loosing data, etc...)

OK?

[Ishan]

Hello Aib

Thank you very much for your reply.. It worked fine for me.

I have few question from you,

1. After i created the Discovery rule, item prototype & graphs. i had 9 items and 3 graphs by mentioning channel numbers staring from 0-9. Why only three were showing while we do the snmpwalk? (and only three graphs as well)

2. While i was creating the item prototype i have selected the type of information as Numeric (unassigned) data type - decimal and units- bps is that correct? or do i need to select the Numeric (Float)

3. In your system have you been created triggers?

Thanks for your solution

Kind Regards,
Ishan

[aib]

I glad to hear that you have working Discovery rule!

1) Sorry, I cannot explain what happens in your system if I cannot see your source Discovery rule and the Items/Graphs which you have as a result.

2) Numeric(Float) type is for Temperature. Or you believe that your traffic can be negative? I think you was incorrect when typed the word UNSIGNED as a word (unassigned).

3) Yes, I did. But it is only you who can decide how your trigger has to work. What level of load/signal has to "flip the switch". Check the typical Discovery Rules from Templates. Check the Templates on Zabbix Wiki. Some of them has a good triggers which can help you to create your own.

Have fun!

[ouamly]

First of all, not all servers/devices has SNMP daemon working by default.
Then, not all manufacturers like to publish data in SNMP.

And short answer for your question "It is possible to get exactly the sames measures on my aix servers like if i use the aix zabbix agent?" = NO.

We can discuss more details, but as I can see on Zabbix Download Page, you can install Zabbix Agent for AIX 6.1. Why you don't like to do that?

Hello aib!

I thank you for answer. I'm a trainee and my supervisor does not want to install the agent on servers. that's why I have to use snmp. To monitor my AIX server, I created a rule discovery with the OID of the measure, but I still do not have lift. This may be due to what?

Thank you!

[aib]

Did you try to use the standard Template - Template SNMP OS Linux?
It can collect a lot of information, it has 3 auto Discovery rules, and will create several dozen of Items.

Try it, it's free

[ouamly]

Did you try to use the standard Template - Template SNMP OS Linux?
It can collect a lot of information, it has 3 auto Discovery rules, and will create several dozen of Items.

Try it, it's free

Yes, i'm using it but it can discover just DIsk partitions, Network interfaces and processors. I want to get on my AIX server the processes who running, memory utilization, cpu utilization, banwith using by interfaces. thank

[aib]

I use my GOOGLE power and - voila!

The OID I use on aix 5.1 :

For CPU :

OID xmdCPUGluser .1.3.6.1.4.1.2.3.1.2.2.2.1.1.101
OID xmdCPUGlkern .1.3.6.1.4.1.2.3.1.2.2.2.1.1.102
OID xmdCPUGlwait .1.3.6.1.4.1.2.3.1.2.2.2.1.1.103
OID xmdCPUGlidle .1.3.6.1.4.1.2.3.1.2.2.2.1.1.104
OID xmdCPUGluticks .1.3.6.1.4.1.2.3.1.2.2.2.1.1.121
OID xmdCPUGlkticks .1.3.6.1.4.1.2.3.1.2.2.2.1.1.122
OID xmdCPUGlwticks .1.3.6.1.4.1.2.3.1.2.2.2.1.1.123
OID xmdCPUGliticks .1.3.6.1.4.1.2.3.1.2.2.2.1.1.124

For Memory :

OID xmdMemRealEfree .1.3.6.1.4.1.2.3.1.2.2.2.1.2.1.10
OID xmdMemRealEpinned .1.3.6.1.4.1.2.3.1.2.2.2.1.2.1.11
OID xmdMemRealEcomp .1.3.6.1.4.1.2.3.1.2.2.2.1.2.1.12
OID xmdMemRealEnoncomp .1.3.6.1.4.1.2.3.1.2.2.2.1.2.1.13
OID xmdMemRealElocal .1.3.6.1.4.1.2.3.1.2.2.2.1.2.1.14
OID xmdMemRealEclnt .1.3.6.1.4.1.2.3.1.2.2.2.1.2.1.15

An also IBM published very detailed article Systems Administration Toolkit: Using SNMP data

[ouamly]

I use my GOOGLE power and - voila!

The OID I use on aix 5.1 :

For CPU :

OID xmdCPUGluser .1.3.6.1.4.1.2.3.1.2.2.2.1.1.101
OID xmdCPUGlkern .1.3.6.1.4.1.2.3.1.2.2.2.1.1.102
OID xmdCPUGlwait .1.3.6.1.4.1.2.3.1.2.2.2.1.1.103
OID xmdCPUGlidle .1.3.6.1.4.1.2.3.1.2.2.2.1.1.104
OID xmdCPUGluticks .1.3.6.1.4.1.2.3.1.2.2.2.1.1.121
OID xmdCPUGlkticks .1.3.6.1.4.1.2.3.1.2.2.2.1.1.122
OID xmdCPUGlwticks .1.3.6.1.4.1.2.3.1.2.2.2.1.1.123
OID xmdCPUGliticks .1.3.6.1.4.1.2.3.1.2.2.2.1.1.124

For Memory :

OID xmdMemRealEfree .1.3.6.1.4.1.2.3.1.2.2.2.1.2.1.10
OID xmdMemRealEpinned .1.3.6.1.4.1.2.3.1.2.2.2.1.2.1.11
OID xmdMemRealEcomp .1.3.6.1.4.1.2.3.1.2.2.2.1.2.1.12
OID xmdMemRealEnoncomp .1.3.6.1.4.1.2.3.1.2.2.2.1.2.1.13
OID xmdMemRealElocal .1.3.6.1.4.1.2.3.1.2.2.2.1.2.1.14
OID xmdMemRealEclnt .1.3.6.1.4.1.2.3.1.2.2.2.1.2.1.15

An also IBM published very detailed article Systems Administration Toolkit: Using SNMP data

I work on aix 6.1 and 5.3. do you think that OIDs are the same! I will look at the link you sent me on IBM and do some tests. I'll let you know. Thank you very much.

[aib]

I work on aix 6.1 and 5.3. do you think that OIDs are the same!

Yes, I do.
Usually old OID works for any version of OS.
Except the situation when OID was especially added for the new system.

I will look at the link you sent me on IBM and do some tests. I'll let you know. Thank you very much.

Good luck

[ouamly]

Yes, I do.
Usually old OID works for any version of OS.
Except the situation when OID was especially added for the new system.



Good luck

Thank you!

[Ishan]

Hello Aib

Thanks for your advise, i will be attaching my discovery rule, Item prototype, graphs screenshot and items which was created automatically. Can you kindly advise me how to make triggers like,
1.While there is no traffic on the PRI (to monitor ISP disconnection)
2. If there is a disconnection on PRI
3. If PRI administratively disconnection
4. If PRI usage high traffic

And i would like to clarify with, why the PRI splitting to 3 channels and giving different values of traffic on each?

Thanks & Kind Regards,
Ishan

[aib]

Thanks for your advise, i will be attaching my discovery rule, Item prototype, graphs screenshot and items which was created automatically. Can you kindly advise me how to make triggers like,
1.While there is no traffic on the PRI (to monitor ISP disconnection)
2. If there is a disconnection on PRI
3. If PRI administratively disconnection
4. If PRI usage high traffic

1) Create a trigger
{channel.load["{#SNMPVALUE}"].avg(300)}=0
(average load for last 5 minutes equal 0)

2) you can use the same trigger or create one more for 10-15-20 minutes of no traffic.

3) For checking the administrative status you have to create one more Item per channel which contain "Status" of channel. After that you can use it in Triggers.

4) Create a trigger
{channel.load["{#SNMPVALUE}"].avg(300)}>130000
(average load for last 5 minutes more than 130000 bytes per second = almost 128 kBps)
Use different numbers which you think is more fit for your needs.

And i would like to clarify with, why the PRI splitting to 3 channels and giving different values of traffic on each?

I have no idea. May be some manufacturers documentation can explain something?

[Ishan]

Hello Aib

Thanks for your reply and i will try that on after few hours.

I have some other issue with graphs now, it's frequently disconnecting, I will attached the screenshot now, please advice me what i can try with that.Yesterday and today it was disconnected and sometime i need to restart the zabbix-server to get the data the graphs, sometimes it's coming back automatically.

[Ishan]

Hello Aib

Sorry here is the screenshot