Dear Network Administrators,

This CPTTM NetAdmin newsletter is to bring useful news to you, Network Administrators in Macau, for references without obligations, so that you can do your jobs easier and better! Hope you like it. if you'd like to unsubscribe or recommend your friends to subscribe, just email me at kent@cpttm.org.mo. Old issues are available here. Also printable version for this newsletter.

Topics in this issue:

• Retaining mails in MS Exchange mail boxes
  
• Setting up a high availability service is easy & free
• Testing network speeds
• Prevening time drifts in VM guests
• Upcoming courses for network administrators

Retaining mails in MS Exchange mail boxes

A very useful feature in MS Exchange 2007 is that you can let users file their mails into different folders (e.g., Personal, R&D and Legal) and then retain the mails in those folders for different periods, as required by legal or administrative requirements (e.g., keep mails in the Personal folder for 1 month and those in R&D for 5 years). This is more flexible than keeping every single mail for the same period (or forever) as there are sometimes company policies requiring deletion of mails.

To learn more about this and other features of MS Exchange 2007, please join our up-coming Microsoft Official Course: Introduction to Installing and Managing Microsoft Exchange Server 2007 (5047A).

Setting up a high availability service is easy & free

Possible to make sure a certain IT service (e.g., mail) available even if the server is down, say, due to hardware problems? You may think that this may cost a lot of money and a lot of work, while in fact, it can be easy and free!

At CPTTM  we have set up a cluster for mails (SMTP, IMAP, POP3, web mail) and it took only a couple of days and with no licensing cost at all.

How does it work? As shown in the figure, host 1 is hosting some VMs that provide the actual services (SMTP, web mail, etc.). Normally the VMs are run on host 1 only. This is a very normal setup. To make sure the services are still available even if host 1 is down, the idea is to let host 2 monitor the health of host 1. If it is down, host 2 should run the VMs on itself. This "monitor and respond" action is done by a software package called Heartbeat. How to make sure the VM images are available on host 2? You could put the VM images in a shared storage such as SAN or NAS. For us, we replicate the disk partition on host 1 to host 2 using the DRBD device in Linux. To ensure fast replication, we use a Gigabite Ethernet crossover cable to link up the two hosts. 

This setup is very simple and costs nothing, but works very well. I've tried shutting down host 1 and then the web mail did become available again in a few seconds.

You may wonder as host 2 is usually idle, we may be wasting the hardware resource? In fact, we set up each of the host run a certain number of VMs and monitor each other. If the other is down, it will run the VMs originally run by the other (and thus will become slower).

For step-by-step instructions, you may follow this excellent tutorial.

Testing network speeds

When users complain that an IT service is slow, it is difficult to find out which part is slow: the application, the network or the database? One thing the networking team can do is to test the network speed. If it is fast, it means it is not their fault. If it is slow, it means there is a problem somewhere on the network. How to test the speed? There is a simple and free tool called iperf. It has both Linux and Windows versions. Here is a tutorial on it. We successfully used it to identify a network bottleneck caused by mis-configuration of a firewall and another one caused by a native design fault in a switch.

Preventing time drifts in VM guests

It is very common for the clocks in VM guests to drift away, as the hardware clock tick just can't be delivered to the VM guest OS if that VM is not running at the moment. How to alleviate the problem? If the guest is Windows, install VMWare tools and enable "periodic synchronization" in it. If it is Linux, follow the best pratices. For the glory details of how these things work, click here.