Installing APC PowerChute Network Shutdown On A VMware vSphere 4 Host
I did this today for a client of ours because they did not have it installed and the local power company will be shutting them down on Sunday for maintenance. APC PowerChute Network Shutdown is used when your APC UPS has one of their network cards installed in it. The Windows version is free, however the VMware version is not, it costs $99 which isn't bad but personally I think it should be free as well. You can purchase the download here, however be sure sure to check their compatibility matrix to make sure your system is supported.
After buying the software you can download it immediately, I downloaded it on my vCenter server. Once I had it downloaded I opened the zip file and found multiple folders for different operating systems: ESX, ESXi, Windows, and Server 2008 Core. This client is running VMware vSphere 4 so I will be using the file in the ESX folder.
Open your vSphere Client and browse to a datatore and upload the pcns224esx.tar.gz file
SSH into the host and browse to your datastore which can be local or on shared storage.
cd /vmfs/volumes/<GUID>
Move the file to your home directory.
mv /home/<username>
Extract the compressed file.
tar -zxvf pcns224esx.tar.gz
Start the installation.
./install.sh
Press any key to view the license terms and space bar until you are done then type "yes" to agree to them. If you have only one UPS select 1 for the number of instances, if you have more select the appropriate number. I allowed the software to use the default installation directory and the bundled JRE which is version 1.5 update 18.
------------------------------------------------------------------
PowerChute Network Shutdown v.2.2.4 for VMware
Copyright (c) 2009 Schneider Electric, All Rights Reserved.
------------------------------------------------------------------OS=Linux
Initializing ...
Press any key to display End User License Agreement
Do you agree to the above license terms? [yes or no]
yesPlease enter the PCNS instance number [1|2|3] or press enter to use default value of 1:
1 PCNS instance(s) will be installed.
Please enter the installation directory or press enter to install to the default directory (/opt/APC/PowerChute):
Are you sure you want to install PCNS to /opt/APC/PowerChute [Yes|No]?
Yes
Creating /opt/APC directory ...
PCNS will be installed to /opt/APC/PowerChute
Copying the installation files ...
Extracting PCNS files ...
PCNS is extracted to /opt/APC/PowerChutePlease enter java directory if you want to use your system java (example:/usr/local/bin/jre/jre1.5.0_18) or press enter to install the bundled Java:
Copying jre to /opt/APC/PowerChute/jre ...
Extracting jre to /opt/APC/PowerChute/jre ...java version "1.5.0_18"
Java(TM) 2 Runtime Environment, Standard Edition (build 1.5.0_18-b02)
Java HotSpot(TM) Client VM (build 1.5.0_18-b02, mixed mode, sharing)JAVA_DIR=/opt/APC/PowerChute/jre/jre1.5.0_18
Configuring startup files ...
Startup script=/etc/rc.d/init.d/PowerChute
Updating Linux symbolic link ...
2010-08-21 05:00:57 (28215) WARN : OpenPort: Opening a previously opened port: 3052.
2010-08-21 05:00:59 (28749) WARN : OpenPort: Opening a previously opened port: 3052.
2010-08-21 05:01:00 (29226) WARN : OpenPort: Opening a previously opened port: 3052.
Configuring uninstall script ...Completed.
Please run the PCNSConfig.sh script located within the PCNS installation directory to complete the installation.
After the installation has finished browse to the installation directory.
cd /opt/APC/PowerChute/group1/
Run the configuration file.
./PCNSConfig.sh
It will ask you to select the IP address that you want to register with the UPS. In my case I selected #3 which is the service console of the host, the other two are the iSCSI connections. It will also ask you for your subnet and how many UPS you have and the type.
It will want the IP address of your UPS, the port being used, the username, the password, and the authentication phrase. It did not like the password I entered, as it cannot be empty and no more than ten characters which mine was. I logged into the UPS and changed the password to something shorter that it would like.
If you have not configured an authentication phrase log into your UPS, select the UPS tab, select PowerChute, then configure. Also select the configuration link then the shutdown link and change the low battery duration to 5 minutes.
You should then be able to register with the UPS and start the service. Once you have done that you can browse to https://IP Address:6547 and login with the username/password you used in this configuration file.
--------------------------------------------------
PowerChute Network Shutdown Configuration Utility
--------------------------------------------------Press Ctrl + C at anytime to abort.
[1]: 192.168.131.9
[2]: 192.168.130.9
[3]: 192.168.0.228Please select the local host IP address to register [ 1 - 3 ]: 3
Please enter the subnet mask of the local host IP address to register or just press enter key to use default subnet mask [255.255.255.0]:
Configuring PowerChute Network Shutdown ...
[1]: Configure for a single UPS device
[2]: Configure for a parallel Silcon UPS / Smart-UPS VT / Galaxy system
[3]: Configure for multiple Smart-UPS or Galaxy devices
[4]: Configure for multiple Symmetra devicesPlease select the appropriate configuration type (1) [ 1 - 4 ]: 1
Management Card IP: 192.168.0.210
Management Card Port # (80):
Administrator User Name: apc
Administrator Password: [ MASKED ]
Authentication Phrase (default): [MASKED ]Setting Summary:
Management Card IP: 192.168.0.210
Management Card Port #: 80
Administrator User Name: apc
Administrator Password: [ MASKED ]
Authentication Phrase (default): [ MASKED ]Do you wish to register these settings [ Yes | No | Abort ]? Yes
Registering PowerChute Network Shutdown with the management card ...
PowerChute Network Shutdown registration completed successfully.Do you wish to start the PowerChute Network Shutdown service [ Yes | No ]? Yes
PowerChute Network Shutdown service started.Configuration completed.
Configuring VMware ESXi 3.5 NIC Teaming
NIC redundancy is easy to do in VMware ESXi and should be done if your system has two NICs and if your system only has one I would recommend purchasing a second one. Once you have installed VMware ESXi you will be able to make minimal configuration changes. With a fresh install be sure to change the root password which is the first option, after that go into Configure Management Network.
By default only the primary NIC will be used, the right pane shows vmnic0 being used, to create the team go into Network Adapters.
Select each NIC that you want to include in the team, you will press space bar to select them, for this installation I am going to use only two of the four NICs. When you are done press enter and it will take you back to the previous screen where you should see vmnic0 and vmnic1 on the right pane.
Since your primary NIC probably picked up an IP address from DHCP you will want to change it to a static, to do this go into IP configuration and select "Set static IP address and network configuration".
Verify that your DNS settings are correct if they were picked up by DHCP, if not change them to suite your network. You would also change your hostname in this section, press enter when done to go back to the previous screen.
When you exit back to the main screen it will prompt you to save the changes and restart the management network, press Y to do this and you should be good to go. With any redundant system it is always wise to verify that it is working as intended. Now would be the ideal time since the system is not in production yet, start a continuous ping and unplug the primary NIC, you should only see a few packets drop if any. Plug the primary NIC back in, wait a few seconds and unplug the secondary NIC and your ping should continue.
How Mouse Poop Killed our VMware vCenter Server
With all the redundancy built into a server these days apparently a mouse is able to bring down a server. We had our VMware VirtualCenter running on an old Dell PowerEdge 2650 and the other day I was told it just wouldn't boot. This morning I went about troubleshooting, the LCD on the front said that there was no memory so I opened up the case to make sure the memory was seated correctly. As soon as I opened it I noticed what looked like rust on what I believe is the northbridge and other areas of the motherboard. Upon further investigation I noticed that it looked like mouse poop. Normally this probably would have seemed ridiculous but we were aware that there was a mouse in the office since an employee had seen it and we had noticed that it had been eating our food and pooping. The mouse seems to have gained entry into the server through one of the PCI slots that did not have a cover on it.
We have an extra working Dell PowerEdge 2650 so I put the two hard drives which were configured with RAID and thankfully the server booted and all is well. So, I guess a lesson learned is in addition to server redundancy, backups, and all the other fancy nice things that go along with a stable network infrastructure you need to make sure you don't have rodents in your server room, or at least make sure your PCI slots are covered.
Partitioning VMware ESX 3.5
I recently started working with VMware ESX 3.5 and did my first implementation for a medium size company. Since I was/am no expert in VMware I decided to call VMware support to get their recommendations on partion configurations. Since VMware ESX 3.5 uses very little space for the installation two small (73GB) 15,000RPM SAS drives in a RAID 1 set are ideal. VMware suggested using 10GB for / 10GB for /var 1GB for swap and the remaining space for a vmfs3 partiton to store data on. Because virtual machines will reside on the SAN this data partition could be used for things like ISO images or other data.
