monitoring

[bobby_31]

I am running redhat 7.3 on DL 380 with smart array 5i controller the raid 5 disks are in external compaq storage works 4300 box , Please advise me how can i monitor the disks so that if any disk fails or any thing wrong with the logical drive it should inform me. I tried removing a disk from the array and the driver reported nothing in proc or syslog or dmesg. even the compaq 4300 didn't made any beep , it just turned a small red light on .

Cheers

Bobby

[bobby_31]

Thanks for your reply, Well, you can configure the syslog.conf to get the messages at the specified place. Going a level below i.e. how the message is generated in the first instance . The smart array 5i card should trigger the drive failure first. If we talk about storageworks 4300 . The fault bus signal (SWAP) detects the removal or insertion of storeage device on a SCSI bus , causing the signal to generate a logic low pulse. This pulse is detected by EMU and an amber light is glown . unfortunately there is no beep sound. That was my first question is it done deleberately (design) or something broken in my hardware.

syslog will only get message if the driver will generate anything that was my second question why the cciss does not generate any error when the drive is removed . I guess there is some sort of monitoring utility (SNMP or something that logs the error conditon in proc or syslog) that Compaq provides. Getting that utility was the ultimate goal of this posting .

cheers

Bobby

[Jared Middleton]

Bobby,

You can find the latest HP/Compaq software and drivers for a DL380 running RH Linux 7.3 here:
http://h18007.www1.hp.com/support/files/server/us/locate/86_4706.html

Have you installed the ProLiant Support Pack? The package includes "health" agents that monitor various subsystems:

http://h18007.www1.hp.com/support/files/server/us/download/10931.html

-Jared

[Jarle Bjorgeengen]

Install the Proliant management agents, and health drivers. Then you can set up your favorite management software to receive SNMP traps for hardware events generated by your management agents.

Rgds Jarle