ilo2 to send snmp traps about hardware (raid, power, etc..) without any OS support

[john fish]

Hi.

Server: proliant dl360 g6.

I can't figure out wheter ILO2 is capable to send SNMP traps about critical HW issues (like sata failure, temperatures, power failuer etc)..



I see that ILO2 can monitor this.
Testing snmp traps works, i.e. I receive trap to my remove snmptrapd. But only traps I receive now are test traps and traps on reboot.

Nothing in case I plug out one of power cables or hotswap disks.

I have tested those events with PSP , it can send SNMP traps.

But well what I really want is to have ILO2 working completley independetly, and send SNMP traps about those things that ILO2 is capable to monitor.

[David Claypool]

This is not currently a feature of iLO2.

[kris rombauts]

John,

while some of it is possible (see copy below) but for the networking, disks, memory errors etc you do need the Insight Agents active on the OS running on the server.


Enable iLO SNMP Alerts :

Alert conditions detected by iLO independently of the host operating system can be sent to specified SNMP alert destinations such as Insight Manager or pagers. iLO SNMP alerts include major host system events such as remote server power outages or server resets as well as iLO events such as Event Log filled or a failed login attempt.
iLO sends SNMP traps to an Insight Manager console when an alert occurs. The default setting is No.



Forward Insight Manager Agent SNMP Alerts

Alert conditions detected by host management agents can be forwarded to SNMP alert destinations through iLO. These alerts are generated by the Insight Management agents, which are provided for each supported network operating system. The agents must be installed on the host server to receive these alerts.
These alerts are sent to Insight Manager clients on the network and are forwarded asynchronously by iLO to the IP addresses that have been configured to receive them. The default setting is No.


Kris

[john fish]

Hi,

Thank you for replies.

Does anyknow have good experience about how stable insight / agents / kernel modules for ilo hpasm etc.. are?
I am interested in linux, RHEL5 and CentOS5 (x86_64)

I've read some threads at least about hpasm being not recommended for production use.

Now I've tested rhel5_x86_64 in last days and experienced some weird reboots. I don't know if its because of insight manager,
but anyway I've worried about using insight/agents/kernelmodules in production servers.

[john fish]

What would be alternatives for monitoring most important hardware with linux + proliant dl360 g6.

I would like to use some lightweight and stable opensource solutions.
Just to send snmp/email/nagios when raid/cpu use/etc.. goes critical.

Can anyone recommend something ?
Thanks.

[john fish]

is it possible to query HW status from ilo somehow, xml or so?

[David Claypool · Accepted solution]

@john fish: The information you're looking at is old. The HP health package (hpasm) operates in user space, not kernel space and has for several years now.

[john fish]

OK, thanks.

But I see that for ILO , with ssh I can get pretty much all information.
a'la
bash$ ssh ilo.ip -ladminilo show system1/drives

i could get this work for me with some scripting, but i'm curious if there already way to get for example XML from ilo?

currently i would have to do this with ssh no-passphrase key + parse ilo text output, it's kinda hacking.

[Krzysztof Bak]

Hi

I set all enable in iLO2 > Administration > Management and i haven't catch any trap except test trap.

I testing by unplug one of power cord.

I check this in

iLO2 > System Status > IML

and I see

System Power Supplies Not Redundant

but my SNMP server don't catch any trap

I have HP Proliant 380 G6
SNMP server is Nagios

is anybody know where is the problem?

[ThomasHansen]

We are experiencing the exact same thing on a ML380 G6.

Did you ever get this fixed?