As its in test, it got stuck on our currently relatively unused “old” vmware environment (ESX 3). After initially going well it start to go wrong. We kept getting critical alerts that a server was up. Switching zenping to debug didn’t help either – it offered no new information and made the problem worse. As its open source I thought I’d take a look at the source, and hey presto I found the suspected problem. Stuck in an extra debug line, and confirmed it.

The problem – well it was an issue with the clock on the server caused by running under vmware , it was jumping about leading to negative rrt on pings. Funnily enough zenoss didn’t like. I’ll be submitting a bug so that it comes up with the slightly less cryptic error of Ip xxx.xxx.xxx.xx is up. To fix the problem, I had to specify the clocksource in the kernel options. See vmware KB 1006427 for details.

If you can’t fix this for some reason or can’t reboot server, for now you can put an event transform for status/ping in place to suppress them.

import re
 
match = re.search('ip (25[0-5]|2[0-4][0-9]|[01]?[0-9][0-9]?)\.(25[0-5]|2[0-4][0-9]|[01]?[0-9][0-9]?)\.(25[0-5]|2[0-4][0-9]|[01]?[0-9][0-9]?)\.(25[0-5]|2[0-4][0-9]|[01]?[0-9][0-9]?) is up', evt.message)
if match and evt.severity==5:
	evt._action = 'drop'

special_remote:
  debug_print = "R: special_remote for $local_part@$domain"
  driver = dnslookup
  domains = pobice.com
  local_parts = lsearch;/etc/exim4/non_local
  transport = remote_smtp
  ignore_target_hosts = 127.0.0.0/8
  no_more

I then have entered the local_parts into a file.  Everything else then gets bounced, rather than been forwarded and bounced later on.

Word of warning – I’m no exim expert, but this seems to do the trick for me.

Ubuntu lists it as:

Zeroconf LIRC daemon using input event devices
This is a small LIRC-compatible daemon that reads from /dev/input/eventX
devices and sends the received keycodes to connecting LIRC clients. Inputlircd
needs no configuration, it uses the standardised names for the keycodes as
used by the kernel. Many USB remote controls that present HID devices, as well
as multimedia keyboards should work out of the box.

Why is this good, well I just install this package and well I get full remote support – once I’ve configured the applications, which is still a pain in the arse. Now if the applications could be written to work with the zeroconfig lirc with well zero config the world will be great (well ok less un-great).

Its now running on Debian for the next Cacti box (assuming it doesn’t have the gd bug).

What has helped with the new Cacti box is a combination of a few things – the new 1 min poller patch spreads I/O out better, and allows me to poll device which don’t do 64bit SNMP counters yet have gigabit ethernet ports every min, and leave everything else at 5 min. Improvements on plugins and a switch to cactid (a c version of the poller) has also really helped. The system now process ~400 hots in 35 seconds, dealing with ~4100 data sources and ~2200 rrd each min. Not bad.

There are a couple things that need improving however:
1) The threshold system now inspects data straight from the db before writing to the rrd files, which unfortunately causes the graphs in the email to be 1 cycle out of date (ie not show the data the emails about).
2) The Uptime+Threshold system host down email only works by patching files yourself at present when using cactid

As for ubuntu – either stop shipping cactid, or compile it again the correct version of the libraries. I’d much rather the package not be there than for it to not work, it gives people a bad impression.

For this to work you have to:

1. Have a CPU that support Frequency Scaling
2. Have CPU Frequency scaling working
3. Use NUT (Network UPS Tools)

Here what you need to do to get thing working

• Check you what ‘Governors’ you have available
  You need to make sure the correct modules are loaded so that you can scale to the correct settings. To see what you do have loaded run:
  cat /sys/devices/system/cpu/cpu0/cpufreq/scaling_available_governors
  You’ll need at least powersave and one other. Powersave should set the CPU Frequency to the lowest setting. If you don’t have powersave you’ll need to load the module. As root run:
  modprobe cpufreq_powersave and add cpufreq_powersave to the file
  /etc/modules.

  To test that it is working, ideally run something that force you CPU to 100 for a while (say compress a large folder) and then run
  cpufreq-selector --governor powersave
sleep 1
cat /sys/devices/system/cpu/cpu0/cpufreq/cpuinfo_cur_freq
cpufreq-selector --governor userspace
sleep 1
cat /sys/devices/system/cpu/cpu0/cpufreq/cpuinfo_cur_freq

  The first number given should be the lowest CPU frequency and the last should be the highest (assuming the process is using a lot of CPU time still)

• Add a few new items to upssched.conf
  Using you’re favorite editor open up /etc/nut/upssched.conf and add the following line:
  AT ONBATT * START-TIMER powersave 30
AT ONLINE * CANCEL-TIMER powersave
AT ONLINE * START-TIMER userspace 30
AT ONBATT * CANCEL-TIMER userspace
  I’ve used a timer (30 seconds) to smooth out jumpy power from triggering off any actions.
• Add the command to the CMDSCRIPT file to change the CPU frequency govenor

Edit the cmdscript file (the path is given in upssched.conf on the line starting CMDSCRIPT. This file probably doesn’t exist – so just create a new one. Assuming it doesn’t exist add the following lines:
#! /bin/sh
case $1 in
powersave)
/usr/bin/cpufreq-selector --governor powersave && logger -t upssched-cmd "Setting CPU Governor to powersave"
;;
userspace)
/usr/bin/cpufreq-selector --governor userspace && logger -t upssched-cmd "Setting CPU Governor to userspace"
;;
*)
logger -t upssched-cmd "Unrecognized command: $1"
;;
esac

• Test Unplug you’re UPS for 1 min, plug back in and look at /var/log/syslog – you should see that NUTs has throttled you’re CPU.

I haven’t had a power cut since I’ve implemented this, but from looking at the stats from the UPS I should gain another 7 min or so of battery time – an extra 21%. An improvement on this may be stopping some process and task when on battery and starting them again when back on mains. But for now, at least untill the next power cut this will do for me.

[WUFTPD]
# Option: enabled
# Notes.: enable monitoring for this section.
# Values: [true | false] Default: true
#
enabled = true

# Option: logfile
# Notes.: logfile to monitor.
# Values: FILE Default: /var/log/auth.log
#
logfile = /var/log/auth.log

# Option: port
# Notes.: specifies port to monitor
# Values: [ NUM | STRING ] Default:
#
port = ftp

# Option: timeregex
# Notes.: regex to match timestamp in SSH logfile. For TAI64N format,
# use timeregex = @[0-9a-f]{24}
# Values: [Mar 7 17:53:28]
# Default: \S{3}\s{1,2}\d{1,2} \d{2}:\d{2}:\d{2}
#
timeregex = \S{3}\s{1,2}\d{1,2} \d{2}:\d{2}:\d{2}

# Option: timepattern
# Notes.: format used in "timeregex" fields definition. Note that '%' must be
# escaped with '%' (see http://rgruet.free.fr/PQR2.3.html#timeModule).
# For TAI64N format, use timepattern = tai64n
# Values: TEXT Default: %%b %%d %%H:%%M:%%S
#
timepattern = %%b %%d %%H:%%M:%%S

# Option: failregex
# Notes.: regex to match the password failures messages in the logfile.
# Values: TEXT Default: (?:Authentication failure|Failed (?:keyboard-interactive/pam|password)) for(?: illegal user)? .* from (?:::f{4,6}:)?(?P\S*)
failregex = wu-ftpd\[\d+\]:\s+\(pam_unix\)\s+authentication failure.* rhost=(?P<host>\S*)

To add it back add the following line to /etc/vim/gvimrc

" Also switch on highlighting the last used search pattern.
if has("syntax") && (&t_Co > 2 || has("gui_running"))
syntax on
set hlsearch
endif

Easy to fix – log on. Goto system > administration and set a default printer. Still you’d have though that when I only have one printer, and its the first printer I added it would be the default, by well default

A bit of messing around – including installing the version from Testing I found out I need a new kernel for WPA.

Of I went and got 2.6.17 and compiled it.

Things of Note:
I had to use the ieee80211 module and the ipw2200 module from the kernel – but they appear to work so now problems.

After that network manager poped up – it found my wireless network. Selected it, a setup box appeared in which I put my WPA pass-phrase in and off it went and connected – all working a lot more reliable that just wpa supplicant on its own. Could be down to the new kernel mind. However it is very easy to use, and well I impressed. It’ll make using my laptop as a laptop a lot easier.

Update
Wow – after all this work and finally wireless is the most stable I’ve ever had it on Linux. In fact its not gone wrong once since this. Even in windows I was having a few problems now and again. Speed is good, stability and great and it just works. I enter a password in at the start and it goes off, picks the right network and passpharse and just sets it up for me.