More obstacles to resume

Customers want to do the deployment of a set of Oracle 11g RAC in a test environment, using a virtual machine, you know, I did all the real scene before. No brother has done this task for me at the moment, so I have to roll up my sleeves and do it myself.

All the energy to do fast consumer goods on-off, a long time no contact with the system, get up relatively unfamiliar. During the deployment process, there are a number of obstacles that are documented here.

Connecting virtual machine instances (PROXMOX)

The client gives a Web-based login to the browser, manages the virtual machine, and then, through Java, pops up a class of VNC windows. Start only to the management interface, the console window does not come out, toss and exchange for a while to know is the browser compatibility issues. The console is not present in the middle of the problem because of the low Java version of the local Windows.

650) this.width=650, "style=" Border-top: #ddd 1px solid;border-right: #ddd 1px solid;background:url ("/e/u261/lang/ Zh-cn/images/localimage.png ") no-repeat center;border-bottom: #ddd 1px solid;border-left: #ddd 1px solid;" Src= "/e/u261 /themes/default/images/spacer.gif "alt=" Spacer.gif "/>650" this.width=650; "title=" proxmox.jpg "style=" height : 322px;width:708px; "src=" http://s3.51cto.com/wyfs02/M01/71/72/wKioL1XQxxmzqMCbAAQSDAlgUxA551.jpg "width=" 1200 " height= "433" alt= "wkiol1xqxxmzqmcbaaqsdalguxa551.jpg"/>

However, after logging in this way, it is inconvenient to use, mainly manifested in:

The copy and paste is not good. In particular, copy the text, and copy the error message out to search.

Uploading files is inconvenient. Normally, the installation of Lrzsz can be used to upload local files RZ.

Keyboard input, sometimes pressing a key will be a long string of characters.

The connection is not stable, it is easy to drop the line.

To resolve such inconvenience, strongly ask the other party to install a VPN. I dial from the local, then SSH and VNC can log on to the server (instance of the virtual machine).

Using a VPN

The other side to help me with the VPN, the dial is OK, also can ssh directly connect each server (virtual machine instance). I am preparing the Oracle RAC installation process, there are some errors, want to take the error message to search engines to find out the network does not pass. Try a few websites, login QQ, etc., are not. From this, the VPN server does not enable forwarding rules, can only connect to the intranet. Ask the other party to the system's permissions (test development environment), login up, confirmed, installed is pptpd. And then wrote a script, run it, it's normal. The script reads as follows:

#!/bin/bash #writed by sery,2015-8-15 /sbin/iptables-t nat-a postrouting-s 192.168.40.0/24-o eth3-j SNAT--to-source 121.20.48.132 /sbin/iptables-t nat-a postrouting-s 192.168.42.0/24-o vmbr0-j SNAT--to-source 192.168.40.21 /sbin/iptables-t nat-a postrouting-o Eth3-j Masquerade

I look at the history command, see there is also iptables, but he used the wrong interface, the correct interface name is Vmbr0, did the bridge, not eth1. It took some time to find the problem carefully.

Operating system version

When installing the Oracleasm component, it was found that the system version was too high and did not match Oracle 11g. They had to change the version to cents. 5.11. The kernel version is the official site of the Linux rac101 2.6.18-398.el5,oracle, just to find the package "oracleasm-2.6.18-398.el5-2.0.5-1.el5.x86 _64.rpm ". This is critical, otherwise creating an ASM disk behind will fail.

SSH authentication

Grid installation process, with the installation interface directly click "Setup", success, but then down, will not pass, prompting the ssh between the nodes failed. Do a few more attempts:

Check each node/etc/hosts file, for fear of which character is wrong, and then proceed, invalid.

Manually execute Ssh-keyge, merge two node key files, then manually ssh, two-way login, no password required. Then go back to the grid installation interface to verify or not.

View the logs and find no useful information.

Execute it./runcluvfy.sh stage-post hwos-n Rac101,rac102-verbose, the problem is that another network (the private network of Oracle RAC) does not work. The Mutual ping Intranet address (192.168.41.*) does not pass. Had to harass each other and let them deal with it. Get feedback is to do the host host Nic Bonding, after the cancellation, it is normal.

ASM disk is missing

Installation continues, and when the ASM Disk group is selected, it is found to be empty inside. Re-login system, Cd/dev/oracleasm/disks, inside is also empty. Should be iSCSI failure, notify the other side processing, normal rear can be carried out behind the operation.

ASM instance does not start

After the grid is installed, the ASM Disk group is missing when the Oracle database is installed and the storage location is selected. By viewing it, it is found that the ASM instance of Node 2 cannot be started, and rebooting the system is not possible. A check, found that made a low-level error: ASM instance SID is actually the same as the first SID (generated automatically by the same script, forget to manually modify). Modifying the SID is not a good thing, but deleting the entire grid. A hurry, the execution order is wrong, incredibly deleted. Toss for a while, all ready for them to re-install the system. Take a break and kill the entire installation directory/u01/app. Come back again and be normal at last.

This article is from the "Sery" blog, make sure to keep this source http://sery.blog.51cto.com/10037/1685119

More obstacles to resume