Ethereum multi machine deployment

1. vim vi /etc/sysconfig/network-scripts/ifcfg-eth1
DEVICE=eth1
BOOTPROTO=static
IPADDR=192.168.1.1
NETMASK=255.255.255.0
GATEWAY=你的网关
ONBOOT=yes
vim vi /etc/sysconfig/network-scripts/ifcfg-eth0
DEVICE=eth0
BOOTPROTO=static
IPADDR=10.1.1.1
NETMASK=255.0.0.0
GATEWAY=你的网关
ONBOOT=yes

2. No, there are two modes of logstash, one is server, the other is client. The client takes the current host as the log input, which needs to be installed on each host. However, the server mode does not. It requests the remote host to input the log, that is, when the Java program finds that logstash is connected, it will input the log into logstash

3. The deployment problem can be solved by fabric or ansible. To increase the workload of a machine is to change a configuration item
define the deployment workflow in the deployment script, which is generally:

prepare the environment, installation depends on
update code
build & compile
restart the service
for the event that the monitoring application hangs up and restarts automatically, you can give it to the supervisor: a process control system

4. So what you said downstairs is right. You can only call the WebService on each machine separately. For the so-called configuration, just write a list of URLs. The WS client circulates to call the address in the list and saves the data.

5. 1. PXE + kickstart + DHCP + NFS unattended Linux Installation, the premise is that the computer you want to install supports PXE startup< All network cards need to support PXE startup,
2. Install a kickstart server, DHCP server, TFTP server (the above can be installed on the same server), configure kickstart configuration,
3. Other servers are configured for PXE network card startup, after startup, IP can be automatically assigned, and system installation can be started by getting kickstart configuration.

6. According to my understanding of cluster, there are only three cluster modes at present: high availability cluster, load balancing, high performance computing
ha (high availability) cluster
is a cluster that normally works as a running server and transfers business to standby server when the running server fails. It's a cluster for high availability. Including shared disk and mirror disk< br />􀂋 Load balancing cluster
is a cluster that distributes client requests to each node according to the proper load balancing principle. It is a cluster with the purpose of high scalability, and generally can not carry out data handover. It includes load balance cluster and parallel database cluster< br />􀂋 HPC (High Performance Computing) cluster
refers to the cluster with large amount of computation. It's a cluster that uses supercomputers to perform a single business. In recent years, the grid computing technology that uses all nodes
CPU to perform a single service has become a hot spot
note: this paragraph is extracted from the NEC expresscluster proct start guide

according to your needs, it seems that high availability cluster can not achieve multi server parallel at present. The main procts are (NEC expresscluster, Veritas of Symantec, autostart of EMC and some domestic procts), and load balancing can realize load sharing, However, although the current load balancing realizes the business sharing to each server, it can't realize the data synchronization. (representative procts include F5, NEC's iplb and Barracuda)
as for high-performance computing, I haven't had much contact with it, so I may need to have in-depth communication according to the needs of the owner!

7. Or you can separate out a VLAN on the layer 3 switch to the server, and then set the rule that you can't access the external network. Or you need a Wang Shuai.

9. Server cluster: a server cluster is to gather many servers together for the same service, which seems to the client as if there is only one server. Cluster can use multiple computers for parallel computing, so as to obtain high computing speed, and can also use multiple computers for backup, so that any machine is broken, and the whole system can be recovered

10. Deploying multiple services, is it the same service? There should be configuration files that can be changed to monitor ports, so that they can be separated