topic Re: hpacucli to much logs in /var/log/messages in Operating System - Linux

hpacucli to much logs in /var/log/messages

karel barel — Thu, 30 Sep 2010 18:49:15 GMT

My monitoring runs hpacucli little bit frequently, each time are writen to /var/log/messages these lines

Sep 30 21:30:04 hp214 kernel: blocks= 429925920 block_size= 512
Sep 30 21:30:04 hp214 kernel: heads= 255, sectors= 32, cylinders= 52687
Sep 30 21:30:04 hp214 kernel:

Why hpacucli sends these logs?
Is posible disable it?

hpacucli-8.50-6.0 but same with older releases.

Re: hpacucli to much logs in /var/log/messages

Matti_Kurkela — Sat, 02 Oct 2010 06:18:12 GMT

This message is sent by the kernel, not by hpacucli.

Apparently your monitoring runs some hpacucli command that causes the kernel to re-read the disk geometry.

Which hpacucli command(s) are you running, exactly?

At least on RHEL 5, you can use the smartctl command to query SmartArray disk state instead of using hpacucli. Unlike hpacucli, smartctl won't seem to cause disk geometry re-reads.

Example:
If /dev/cciss/c0d0 is a mirrored system disk and the mirror has 2 physical dsks (= simple RAID1), then you can query the state of the first physical disk with:

smartctl -a -d cciss,0 /dev/cciss/c0d0

and the second physical disk with:

smartctl -a -d cciss,1 /dev/cciss/c0d0

MK

Re: hpacucli to much logs in /var/log/messages

karel barel — Sun, 03 Oct 2010 18:45:07 GMT

each of this commands causes log messeage
hpacucli controller slot=$slot show
hpacucli controller slot=$slot array $array show
hpacucli controller slot=$slot physicaldrive all show

You are right about smartctl, but that doesn't provide infos about raid array status.

I'm surprised why show command re-reads the disk geometry.

Re: hpacucli to much logs in /var/log/messages

Matti_Kurkela — Tue, 05 Oct 2010 13:11:49 GMT

I'm surprised about that too... perhaps the designer of hpacucli thought it was going to be used as a configuration tool only, not for monitoring.

If your monitoring can use SNMP, the hp-snmp-agents will certainly offer a better interface for monitoring.

Some versions of the Linux SmartArray driver offer a bit of information about the status of the array in /proc/driver/cciss/cciss* files.

What's the name and version of your Linux distribution, and what kind of monitoring solution do you use? Perhaps someone else has already solved this problem for a similar set-up.

MK