Proxmox 1.9 – not fully updated (1.9 build 6542)
Had backups set to backup snapshots of machines every day starting at 2AM. They generally took about 3 hours and would finish by 5AM.

Little did I realize that I was no longer receiving the emails saying “backup successful”. Last one was 4 weeks ago. Whoops doesn’t even start to cover that.

Checking it out it looks like the backup stalled on one of the virtual machines.
ps auxww |grep vzdump
Then just kill any PID that looks like vzdump is running.

Re-run the command to backup either from the GUI or from the command line:
vzdump --quiet --snapshot --storage STORAGE --maxfiles 5 --mailto admin 101 102 103 etc

Ubuntu 10.04.3 LTS x32 fully patched.

I needed to install SNMP so my cacti installation could work with graphing. And I thought I had already installed SNMP on the server I was testing against.

Testing from cacti:
snmpwalk -v 1 -c public IPOFSERVER

Timeout: No Response from IPOFSERVER

OK, so let me see if the process is running:
ps aux | grep snmp

root 439193 0.0 0.0 3884 616 pts/0 S+ 09:59 0:00 grep snmp

Well, looks like I need to install it:
apt-get install snmpd

The installation auto-starts the snmpd service with the default configuration file. I verified it was working with ps aux again, and the snmpd service is running under the snmp user.
On the local machine I could now run snmpwalk:
snmpwalk -v 1 -c public localhost
Which retrieved the information I wanted. Unfortunately I still could not run snmpwalk from the cacti server. Doh!

Edit the snmpd.conf file if you want a different community or RO/RW rights:
nano /etc/snmp/snmpd.conf

Otherwise, just edit the default snmpd configuration:
nano /etc/default/snmpd
Change the following line:

SNMPDOPTS=’-Lsd -Lf /dev/null -u snmp -I -smux -p /var/run/snmpd.pid 127.0.0.1′

To:

SNMPDOPTS=’-Lsd -Lf /dev/null -u snmp -I -smux -p /var/run/snmpd.pid -c /etc/snmp/snmpd.conf’

Save and quit

Restart the snmpd service:
/etc/init.d/snmpd restart

Test again:
snmpwalk -v 1 -c public IPOFSERVER

Profit.

I needed to find an easy, and fairly secure, way to allow a user to login on a website and display their current user’s disk usage. Generally I just navigate to the user’s home directory and run
du -hc
But in this case I needed it to work from the web.

By default, Ubuntu runs apache as www-data. Enter sudoers. My editor’s file is nano.
visudo
Add the following to the bottom of the file:
www-data ALL=NOPASSWD: /usr/bin/du, /usr/bin/ls, other/command
Save and quit

Now your www-data user can run those commands by running sudo command. And it will not allow the user to sudo any other commands.

My PHP page is as follows:
echo exec("sudo du -hc /mnt/ | grep " . $_SESSION["valid_user"] . "| awk '{print $1}'") . " Disk Space Used
";
Obviously the SESSION variable comes from a login page, but it’s a pretty good start for me.

I know there is a GUI for RHEL, but I still prefer to do these things via a command line.

RHEL Setup NTPd

To simply set the time to the time server (correct time one-off), use the following command:
/usr/sbin/ntpdate IP_or_ADDR_of_NTP_Server

In my case it was this command and this result:

-bash-3.2# /usr/sbin/ntpdate 10.4.0.253
18 Jan 08:16:02 ntpdate[13347]: step time server 10.4.0.253 offset 756.199630 sec

As you can see, this server was off by over 12 minutes!

If you want to edit the NTP Servers List:
nano (or vi) /etc/ntpd.conf
Add the specific time server you require
/sbin/service ntpd restart

I recently installed FreeNAS 8.x release and wanted to setup a per user quota. There is no way to set this up using the GUI front end – you can only setup a mount point quota at this time.

To setup a per user quota using the CLI, just login as root using your admin password, then issue the following commands:

zfs set userquota@USERNAME=10g SHARENAME
So in my case I had a /mnt/files filesystem, so I used:
zfs set userquota@thelibrarie=10g files

You can list the user quota
zfs list -o userquota@USERNAME SHARENAME

You can remove the user quota
zfs set userquota@USERNAME=none SHARENAME

View the current Input/Output statistics
zpool iostatzpool upgrade -v

View the current compression for all pools
zfs list -o compression

My current infrastructure budget does not allow for vSphere Server, ESXi 4.1 Enterprise, or OpsView Enterprise. They all have one thing in common: they’re more than $0. OK, I know OpsView comes in a community flavor, but still. ESXi hypervisor 4.1 is also free.

So originally I was going to just run SNMP monitoring through Nagios, but I found out that I needed to enable ESXi’s SNMP before I could attempt that. Long story short, ESXi’s SNMP is not worth monitoring. CIM, the supported protocol from VMWare, apparently is the route I needed to go.

So I found this site:
http://www.matthewjwhite.co.uk/blog/2010/02/09/monitoring-esxi-server-health-using-nagiosopsview/
Which gave me the direction I needed.

Log onto your Nagios Server:
apt-get install python-pywbem
cd /usr/local/nagios/libexec
wget http://communities.vmware.com/servlet/JiveServlet/downloadBody/7170-102-7-4233/check_esx_wbem.py
wget http://it.thelibrarie.com/utilities/check_esxi_hardware.py
chmod +x check_esxi_hardware.py

Check by running the following:
./check_esxi_hardware.py IPOFYOURESXISERVER root password VENDOR
You should receive a simple “OK” to verify everything is working

define command{
command_name check_esxi_hardware
command_line $USER1$/check_esxi_hardware.py -H $HOSTADDRESS -U $ARG1$ -P $ARG2$ -V $ARG3$
}

And the check as defined for one of my ESXi servers

define service{
use generic-service
host_name esx01
service_description ESXi Hardware Monitor
check_command check_esxi_hardware!readonlyuser!somepassword!dell
}

Verify it’s working after the check, you should have the green bar saying OK.

I had to edit a little bit as the previous script did not function through nagios although it did work via the command line. Nagios merely reported everything as OK even though I had errors reported on the CLI.