default task "hardware status polling for server" didn't send alert

[xianjun_1]

I am trying to using SIM to manage servers recently. I found one thing seems strange. through SMH I could see a disk have some issue and the status showed by SMH was failed. I didn't get any alert from SIM. But If I restart the snmp serveice of this server, SIM would send alert. From my understanding, SIM should send alert to me when the default task "hardware status polling for server". could someone give me some explanation? Thanks.

[Rob Buxton]

Hardware Polling will pick up when the server is not there.
If there's a hardware poblem on the server (e.g. failed disk) then the agents pick that up and send an SNMP Trap back to the HP SIM Server.
For that to succeed you need the Trap Destination set on the servers.
You can test the set up, from Control Panel run up the HP Management Agents GUI and use the SNMP Trap Send button which should generate a Major Event for the server.

[xianjun_1]

Hi,Rob. Thanks for your kindly response. you mean "hareware status plling for server" task only take effect to the server doesn't have snmp/wmi support(so that the server couldn't proactively tell SIM there is a hardware failure), Am I right? My snmp configuration is good and SIM can receive test trap from this server.

[Rob Buxton · Accepted solution]

Its the agents on the servers that monitor the hardware. They are the ones that will send a trap on failure or some pre-failure conditions. HPSIM will then process those traps as they happen.
They are two separate processes.
If a server crashes or there's a network issue then the agents cannot send anything to HPSIM. That's where the polling task fits in.
It also updates the HW Icon if the HW Agents have reported an issue.

[xianjun_1]

Hi,Rob.Thanks. I have the following scenario: the disk driver already failed before it's managed by SIM.After it's discoveryed by SIM and I didn't get any alert from SIM, I can only check HS and then switch to SMH for more details. but if I restart snmp service I will get the alert. Is this normal?

[Rob Buxton]

Yes that's the intended behaviour.
Alerts are triggered when they happen, but they're also get triggered if the agents get a restart.
So if the disk has already failed you will not get a message from the agents.

It's the same with the threshold settings, it's a one off snmp trap when the threshold is breached. After that you will get them if the agents restart.

[xianjun_1]

Rob. You are really expert in SIM. You mentioned threshold, and I happened to encounter something strange about threshold setting. First I try to set disk threshold by SMH, the disk is used 70%, and if the threshold is set 65%, SIM will send a alert and it's normal. then I set cpu threshold and want SIM alert me when it's usage is more than 80%. After setting the value, I used the tool to make the CPU always busy as 100%. 30 minutes passed I can find the snmp alert sent by foundation agent from windows event viewer (the data collection is set to 2 minutes, from smh's philosophy, 2*15 = 30 minutes the foundation agent will send alert). but can't get alert from SIM. I have tryied this on my 7 proliant windows server, all disk threshold setting were ok, but CPU threshold setting all failed. seems it's not the snmp setting problem, is there any other tricky for cpu threshold setting??Thanks very much.