Ensure server high Availability-disaster recovery (1) configuration process for Web servers

Background:

Our server environment, is the ESX server, which ran dozens of virtual servers. Most of them are virtual Windows servers, and then each application will have several app servers and then install WebSphere to build cluster to support the Web server, both IBM HTTP server and IIS.

The environment is like this, this time for a region of the server cluster to do disaster recovery drill.

So-called disaster recovery, that is, if a region of the server cluster location disaster, earthquake tsunami, how to do? OK, clone all the servers directly and then restore to Area B, and then simulate a complete a-zone environment by configuring the server and Application server (to ensure that the request can be completed through the public ip-->web server-->app server- DB server, and everything works fine). Finally change the DNS settings, the domain name to the new public IP is good. So this time we're working on this, just like the feeling of a fire drill, to make sure that the server cluster does have high availability (the percentage of server downtime is less than how much).

The process of disaster recovery, in simple terms, is the following steps:

1. Clone all virtual servers in area A, and revert to the ESX server in area B.
2. Open firewall port to ensure the connectivity between servers.
3. Sets the application portion of the server.

The first two steps are fixed routines that can be done according to standard practice. The 3rd step, the application of the part, the uncertainty will be a little more.

This article only describes the configuration process for one of the Web servers. In fact, the Web server is more configuration steps, on the one hand to receive requests from the public IP, on the other hand to ensure that the request can be forwarded to the specified app server.

This Web server runs both IIS and many of the IHS (IBM HTTP Server) instances. The purpose of the Web server configuration is to make it clear that the IIS instance and IHS are running and listening to their respective new assigned IP and port, accepting the request and forwarding it to the correct app server.

Configuration process:

1) Set up the network card:

The purpose of this setting is to allow this server to have a physical IP address and multiple virtual IP addresses, and then each IP address can point to a single instance of Web server, including IIS and IHS, so that a single server can accept requests through multiple different IPs. This is a more common practice, because after all, the function of the Web server is to forward the request and processing some static pages, so very little resources, a server has a lot of IHS or Apache instances are very common.

The configuration of Windows Server is simple:

1. Open the network adapter settings, locate the NIC for this server, and then open the properties and double-click TCP/IP V4 settings.
2. You can see the interface of the IP address setting. Of course, a server want to have multiple IP addresses, of course, it is not possible to use DHCP, it must be allowed to have a fixed physical IP address. So just enter the IP address, mask, network management, DNS server.
3. Then click Advanced, the interface that appears, click "Add", directly add virtual IP and mask just fine. Click OK when you're done.

   Verify that simply by pinging the xx.xx.xx.xx (the added virtual IP) is OK. It is normal to be able to ping the pass.

Next Configure the Application Server section:

Before configuring, it is important to note that because this virtual server is cloned, processes that were already running on the previous server may also be brought to this server, causing some unknown problems. So before you configure it, kill all the related processes (in Windows with Task Manager).

1) First configure the IIS server.

The configuration of IIS, normally, is basically all in that configuration management tool. Just right-click on the IIS Web site, click Properties, then click on the "Advanced" button in the graph to get rid of the IIS listening IP address and the corresponding 80 and 443 ports.

2) Next talk about the IHS configuration.

IHS is actually an Apache-based web server software, and many configurations are basically using Apache's configuration method.

For an IHS instance, the most important configuration file is actually two: httpd.conf and Plugin-cfg.xml.

In disaster recovery, the configuration process is as follows:

1. Configure the httpd.conf file: Change all IP addresses to the address you want to listen to.
2. Configuration plugin-cfg.xml file: Ditto ... Of course, it can be regenerated from the WebSphere side and copied here.
3. Launches the corresponding IHS instance. The command is as follows:

<IHS home>/bin/apache-k start-f X:\xxxx\xxx\httpd.conf

Configuration Complete! Then get ready to start and test the service!

Some of the problems encountered during the startup process are left to the next article.