hpacucli causes kernel panic?

[jgeorge0123]

Two of our servers reset. In examining the logs we see that our calls to hpacucli to monitor RAID health take longer and longer to return. It ends with a kernel panic and no kernel logs to the disk.

I see several forum posts on other discussion forums advising against the use of hpacucli for monitoring.

Should we be using hpacucli to monitor RAID health? If not, what is the HP recommended command for checking RAID health on a Linux server in a production environment?

P.S.This thread has been moved from Archived Desktops and Workstations Boards>Linux based to Linux > sysadmin- HP Forums Moderator

[Matti_Kurkela]

The HP-supported way would probably be:

• have snmpd and hp-snmp-agents (included in Proliant Support Pack, also available separately for supported Linux distributions from HP download pages) running
• use the SNMP data (locally or otherwise) to monitor the RAID health: depending on your SmartArray model, you might need CPQFCA-MIB, CPQIDA-MIB and/or CPQSTSYS-MIB to parse the data. The MIB files are available for download here: HP MIB KIT for SIM and other uses

Note that modern versions of snmpd in common enterprise Linux distributions can be configured to listen on 127.0.0.1 only, if you don't need to have it accessible over the network.

If you want to avoid running snmpd entirely, you might look at the CCISS Utilities section of http://cciss.sourceforge.net/

Perhaps cciss_vol_status or cpqarrayd would be appropriate to your requirements.