求助 关于DATASTAGE8.1使用Parallel Job问题
经过几天的努力终于把DATASTAGE8.1安装上了(windows2003),有经过一些时间终于可以进入开发环境了。
但是问题出来了。DATASTAGE8.1上不能使用Parallel Job,搞的我比较郁闷。然后看了一天的安装文档,在安装文档中提到了怎么配置Parallel Job功能。本人E文比较差。有几步搞不明白,希望大家可以给予帮助。
Configuring a computer to be a processing node for MPP
(Windows)
WebSphere DataStage can take advantage of massively parallel processing (MPP).
With a user ID and a secure shell connection, you can configure a WebSphere
DataStage server as a parallel processing node, which can share the processing
load with other WebSphere DataStage servers.
You must have administrator access to the installation system to complete this task.
1. Create a generic WebSphere DataStage user account.
a. Click Start Run cmd to open a Window command prompt. cmd is the
text to enter in the Open: field.
b. From the command window, create the generic dstage user using the
following format:
NET USER dstage password /ADD/COMMENT:"DataStage user account"
/FULLNAME:"DataStage User"/ACTIVE:YES/EXPIRES:NEVER/PASSWORDCHG:NO
password is the password assigned to the dstage user. All machines
participating in the MPP configuration should use the same user and
password.
c. Log off via the Start menu and log in as the dstage user.
d. Launch the Control Panel from Start Control Panel System, click the
Advanced tab, and click Environment variables.
e. Under User variables for dstage, select HOME and click the Edit button. If
the entry contains other environment variables (for example,
%HOMEPATH%) replace the entire contents with the literal value shown in
the User variables list for dstage. The default location is typically
C:\Documents and Settings\dstage. MKS does not have the ability to
resolve these environment variables when connecting remotely.
f. Click OK three times to save your changes and to exit from the System
Applet.
g. Log off as the dstage user and log back in as yourself.
2. Set user permissions. With the new account created, you need to set
permissions on some critical PXEngine directories so that the new user will
have sufficient access rights to execute jobs. The following table lists the
minimum directories that you must set and the appropriate permissions.
Directory Permissions
C:\tmp Write, Read, Modify
C:\IBM\InformationServer\Server\Scratch Write, Read, Modify
C:\IBM\InformationServer\Server\Datasets Write, Read, Modify
a. Using Explorer, browse to one of the directories, right click Sharing and
Security.
b. From the Security tab and under Group or user names, click Add.
c. Enter dstage and click OK. You should see dstage under the list of users
with permissions to this directory.
d. Select dstage and select the permission check boxes for the permissions
Write, Read, and Modify.
e. Click Apply, and then click OK to close the window.
Repeat
these step for each directory that you want to set permissions for.
3. Set remote shell permissions
a. Click Start Run ksh. ksh is the text to enter in the Open: field.
b. From the shell, run the ssh command to connect to your machine. The
format is ssh -l dstage host name. host name is the name of your computer.
You will most likely have to accept an rsa key. You must enter the password
for the dstage user.
c. Once you are connected, run the rsetup command to add the dstage user
and password to the Windows XP/2003 LSA database for use by the rshd
service. Enter the same password when prompted.
d. Edit the .rhosts file in the dstage user!ˉ home directory. You might need to
create the .rhosts file if it does not already exist.
e. Add the line + dstage.
f. Exit from the ssh session.
g. Edit the hosts.equiv file located in the $ROOTDIR/etc folder. The path
name is shown in UNIX format, which is required by MKS. The default
location is C:\Program Files\MKS Toolkit\etc\hosts.equiv.
h. For testing purposes, add the line + + You might need to create this file if it
does not already exist.
i. Test your configuration by running rsh -l dstage host name ls. host name
is the name of your computer. This gives you a listing of the dstage user!ˉ
home directory.
4. Configure PXEngine for RSH/SSH. To run PXEngine parallel jobs on remote
hosts, you need to use a remote shell script. A sample script named
remsh.example is provided for this purpose. The script is located in the
PXEngine\etc folder.
a. Copy the script file, and rename it as remsh.
b. Edit the script file and change the rsh line. Save your changes.
Original script file Changed script file
#!/bin/sh #!/bin/sh
# Example $APT_ORCHHOME/etc/remsh # Example $APT_ORCHHOME/etc/remsh
rsh "$@" rsh -l dstage "$@"
5. Configure the distribute-component script file, which is located in the
PXEngine\etc directory. This script is responsible for managing the installation
and removal of PXEngine component binaries on remote computing nodes.
This script is used when jobs containing compiled PX transforms need to have
their transformer binaries copied to the remote nodes.
a. Make a backup copy of the distribute-component script file.
b. Edit the script file and change the theUser line.
Original script file Changed script file
# The user (optional) # The user (optional
# theUser= # theUser=dstage
c. Save your changes.
You can now run PXEngine parallel jobs in an MPP environment.
这个是全部的说明
其中1.e,3.c,3.d,3.e没有看明白什么意思
希望大家可以踊跃的发言。谢谢。。。。
|