At the beginning of installation, I felt a lot of trouble. Now, I think this process is not too troublesome. Next, I will describe the following in the entire installation process. If there is something wrong or not in place during the entire process, we are welcome to communicate with each other and make common progress.
To install Cloudera Manager, you must first modify the three configurations of the machine:
Disable firewall: service iptables stop;
Disable selinux: setenforce 0 or modify/etc/selinux/config: SELINUX = disabled;
Configure Proxy: Add the following content to/etc/yum. conf: http_proxy = http: // server: port.
It is best to increase the timeout time in/etc/yum. conf, timeout = 55555. Set the length point by yourself.
Prepare the Installation File cloudera-manager-installer.bin, which is a binary file,: Installation. Then set the execute permission, chmod u + x cloudera-manager-installer.bin, and then execute it directly./cloudera-manager-installer.bin. There is a problem here. I cannot run it on my own virtual machine. The problem occurs: cannot execute binary file, probably because it is 32-bit.
During the entire installation process, make sure that your machine is connected to the Internet and you will see a picture of your choice. Generally, you just need to ENTER, next, and accpet, wait until the installation is complete.
Start cloudera manager: enter your host in the browser. My host is http: 192.168.20.195: 7180. Register an account to log in, for example (admin, admin). After Entering the account, we will select the host, and I will install pseudo-distributed, and then install CDH4.1 and impala. This installation process takes a long time, because it is a foreign website, the Network stability has a great impact on the installation. It takes a long time for me to install it. If an installation package error occurs during this period, I must re-start the installation, therefore, it is recommended to use a clean system (without Hadoop-related programs installed). It downloads many rpm packages online. hadoop, hbase, zookeeper, hive, impala and so on are installed under user/lib. Please wait.
After the installation, we can start the services we need. Of course, some services are dependent on each other. cloudera manager is too powerful. It can help you identify and start related services. Shows the service interface I opened:
When the service is enabled, other services are enabled normally, but the most critical services I need are always in bad status, such:
Impala was not started normally, which is the top priority of my research. When something went wrong, I began to look for a problem. Later I found that impala and the database (I used mysql) were not connected successfully. The following describes how to configure the mysql connection impala.
Impala must have hive metastore configured with Mysql or PostgreSQL. The derby database supported by hive is not supported by impala.