HFX micro M2 with MSI IM-GM45

New working slave with less watt usage.

CASE HFX micro M2 - Bundle (EF14 external power supply)
CPU INTEL Core2Quad Q9000 FSB 1066 6M
MOBO MSI IM-GM45
MEM 2x2048MB SO-DIMM Corsair PC2-6400 CL5 KIT
HDD FUJITSU M250 500GB HDD 5400rpm 2,5Z SATA
COOL Coolermaster EPN-41FCS-01-GP (plus custom heatpipe)

Idle ~29WATT (headless linux)
Max ~77WATT (# mprime -t -b4)

New kvm toy received. Totally virtio and managed by virsh from libvirt.

Converting kvm arguments to virsh xml while browsing through libvirt spaghetti code for hints.

shell command

/usr/bin/kvm -boot c -m 1024 -smp 3 -vnc :1 -k en \
-drive index=0,if=virtio,boot=on,file=/data/kvm/kvm-slave.qcow2 \
-net nic,model=virtio -net tap,ifname=tap0,script=no,downscript=no

to virsh xml

<domain type='kvm'>
<name>kvm-slave</name>
<uuid>8974f244-f27b-40f5-a25c-628b91c02e4b</uuid>
<memory>1048576</memory>
<currentMemory>1048576</currentMemory>
<vcpu>3</vcpu>
<os>
<type>hvm</type>
<boot dev='hd'/>
</os>
<features>
<acpi/>
</features>
<clock offset='utc'/>
<on_poweroff>destroy</on_poweroff>
<on_reboot>restart</on_reboot>
<on_crash>destroy</on_crash>
<devices>
<emulator>/usr/bin/kvm</emulator>
<disk type='file' device='disk'>
<source file='/data/kvm/kvm-slave.qcow2,boot=on'/>
<target dev='vda' bus='virtio'/>
</disk>
<interface type='ethernet'>
<target dev='tap0'/>
<model type='virtio'/>
<script path='no'/>
</interface>
<graphics type='vnc' port='5902' listen=':' keymap='en'/>
</devices>
</domain>

Notice:
1) tap0 should not created by virsh so using type=ethernet instead of type=bridge
2) if using disk virtio from file, libvirt 0.6.2 is not adding boot=on, be dirty and using filename as option field (spaghetti code)

SystemRescueCD and IBM TSM backup client

SystemRescueCD is nice
SystemRescueCD is fresh
SystemRescueCD is gentoo
SystemRescueCD is a successor of R.I.P.

Addon script tested with sysrcd 1.1.6 and TSM 5.5.0

#!/bin/bash
#
# SystemRescueCD autorun script for IBM TSM backup client support
#
# function:
# by default SystemRescueCd executes /autorun script if exist
# This autorun script extract archive at SystemRescueCd live filesystem
#
# prerequisite:
# 0) need working server with installed tsm client rpms
# 1) check on working server # rpm -qa | grep TIVsm
# TIVsm-API-5.5.0-4
# TIVsm-BA-5.5.0-4
# 3) create tsm-client.tar.bz2 archive with
# # tar -jcvf tsm-client.tar.bz2 --exclude=/opt/tivoli/tsm/client/ba/bin/tsmstats.* /opt/tivoli/tsm/client
# 4) copy this autorun script at /
# 5) chmod +x /autorun
# 5) copy tsm-client.tar.bz2 at /isolinux/sysrcd-addon/tsm/
# 6) copy your custom dsm.opt and dsm.sys at /isolinux/sysrcd-addon/tsm/
# 7) crosscheck # ls -al /isolinux/sysrcd-addon/tsm/
# -rw-r--r-- 1 root root 856 Dec 19 13:54 dsm.opt
# -rw-r--r-- 1 root root 8774 Dec 19 13:54 dsm.sys
# -rw-r--r-- 1 root root 17948367 Dec 19 15:10 tsm-client.tar.bz2
# 8) optional: example for a isolinux bootcd with sysrcd menu
# # ls -al /isolinux/sysrcd116/
# -r--r--r-- 1 root root 5211568 Feb 24 08:08 altker32
# -r--r--r-- 1 root root 5176624 Feb 24 01:21 altker64
# -r--r--r-- 1 root root 12821248 Feb 25 21:30 initram.igz
# -r--r--r-- 1 root root 3524496 Feb 23 20:40 rescue64
# -r--r--r-- 1 root root 3565712 Feb 25 21:10 rescuecd
# -r--r--r-- 1 root root 212766720 Feb 25 21:19 sysrcd.dat
# -r--r--r-- 1 root root 45 Feb 25 21:30 sysrcd.md5
# # cat /isolinux/rescue_sysrcd.menu
# LABEL 64
# MENU LABEL Kernel for amd64
# KERNEL sysrcd116/rescue64
# APPEND initrd=sysrcd116/initram.igz subdir=isolinux/sysrcd116 backstore=off
#
# history:
# date description
# 01.03.2009 created
##############################################################################
SRCDIR=/mnt/cdrom/isolinux/sysrcd-addon

echo "SystemRescueCd autorun - enable tsm client support"
tar -jxf $SRCDIR/tsm/tsm-client.tar.bz2 -C /
mkdir -p /var/adsmlog
cp $SRCDIR/tsm/dsm.opt /opt/tivoli/tsm/client/ba/bin/
cp $SRCDIR/tsm/dsm.sys /opt/tivoli/tsm/client/ba/bin/
for prog in dsmadmc dsmagent dsmc dsmcad dsmj dsmswitch dsmtca dsmtrace ; do
ln -s ../../opt/tivoli/tsm/client/ba/bin/$prog /usr/bin/$prog
done
for f in libApiDS.so libtsm551xerces-c1_6_0.so libxmlutil-5.5.1.0.so ; do
ln -s ../../opt/tivoli/tsm/client/api/bin/$f /usr/lib/$f
done
for f in libgpfs.so libdmapi.so libha_gs_r.so libct_cu.so ; do
if [ ! -f /usr/lib/$f ] ; then
cp /opt/tivoli/tsm/client/api/bin/$f /usr/lib
fi
done
# EOF

18W LED Lamp (Warm White)

New light? Enough money? Take a future-ready LED lamp.
No, not these cheap ones with only 1W and cool white sick light.

How about this?
18W (9x2W) LED Lamp with 88° beam angle and nice warm white light ( BLT3827D120-XWW92HRL )



comparing with a 50W halo lamp



and both together



Your choose?



Say Yes.

IBM Tivoli Monitoring (ITM) vs nagios nrpe

Your slaves need watching, right?
There is IBM Tivoli Monitoring (ITM) 6.2 packaged as custom rpm

108624525 Bytes itm-6.2-2.6_1.x86_64.rpm

includes

# du -sk *
1112 bin
23044 classes
8152 config
40 etc
7884 gskit
680 HTML
2328 images
1000 InstallITM
72416 JRE
168 keyfiles
804 LAP
920 licenses
136 logs
39884 lx8266
52 misc
100 registry
39280 tmaitm6
8 tmp

and you have nagios nrpe 2.12

36998 Bytes nagios-nrpe-2.12-1.x86_64.rpm

includes

# du -sk *
60 bin
52 etc
2528 libexec
64 libexec-custom
164 share

So, one question: Who survive the wilderness?
The big-fat one or the slim-easy one?

red hat network satellite and older system date

you use also an enterprise installation environment like red hat network satellite and configure channels, kickstart options, fill pre and post script section with custom code. This take some time and at end of a working day you want see some results.
Firing up a new empty server and processing installation fail.
Your are nerved and look at ks-post.log:

# cat ks-post.log
...
An error has occurred:
The SSL certificate failed verification.
See /var/log/up2date for more information
ERROR: unable to read system id.
ERROR: unable to read system id.
ks-post script start
...

So secure connection to your red hat network satellite failed.

Ommm and why?
Maybe you was lazy and forgot to check new server date.
If new server date is older then the satellite SSL cert startdate you will get a SSL error. Comparing with your red hat network satellite

# echo | openssl s_client -connect satellite:443 2>/dev/null | sed -n '/-BEGIN CERTIFICATE-/,/-END CERTIFICATE-/p' | openssl x509 -noout -startdate
notBefore=Dec 24 12:00:00 2008 GMT

you found a reason and smile.

Right and how to fix?
Well to be lazy furthermore, you need a way to say it to your slave.
You can not use kickstart post script (see ks-post.log above) so you will use
the pre script. You know there is ntpdate to correct the new server date, but kickstart pre script environment support only this commands:
http://www.redhat.com/docs/manuals/enterprise/RHEL-5-manual/Installation_Guide-en-US/s1-redhat-config-kickstart-prescript.html
Unfortunately no ntpdate but date and hwclock command.

Set a date after SSL cert startdate and sync it to hardware clock will do the trick at pre script:
# need a server date after satellite SSL cert startdate
# date -s 010100002009
# hwclock --systohc

I'm sure redhat will support in upcoming versions a ntpdate command at pre script section.
In the meantime have a nice day and keep your slaves running...