Thursday, December 9, 2010

Migrating Deployment Manager to other machine.

1 This document outlines necessary steps to migrate the Network Deployment manager (DMGR profile) from one machine to another.

2 The main assumptions before describing the steps are:

3 The IP address of the machine is not referred in any of the configuration or applications.
4 The host name is not hard coded in applications.
5 This document assumes that the new cell installed will have the same “cellname” and ”nodename” value of the source server.
6 Ideal Operating System Platform for DMGR and other server nodes.
7 Identical fix pack levels for source and target servers.
8 The node name at the current environment will not be changed at the time of migration.
9 The cell is not serving the request and it is forcefully made down.


Steps are:

10 Target DMGR Machine:

11 Check for free space. If not create some space to accommodate the compress file containing the DMGR profile catalog.


12 Source DMGR Machine.
13 Check for the empty space on source DMGR Machine.
14 Create a compressed copy of DMGR profile (either tar or zip).
15 Copy the compressed file on target machine.
16 Check the integrity of the zip file copied on the target machine.
17 Get a screen image of the admin console of DMGR for server panel applications panel and SI Bus if existing.


Target DMGR Machine :

18 Unzip the DMGR profile copied from source machine to profiles directory.
19 Make all configuration changes through the deployment manager profile in an ND environment.
20 In the "serverindex.xml" file for the node profile being modified, alter the value of the
21 “hostName” property for the “ServerIndex.xml” element in the file. Search for string “hostName=” to find this property. Also in the “serverindex.xml” file for the node profile being modified, alter the host property for every “EndPoint” in the file. Search all occurrences of “host=” string in following “serverindex.xml” <WAS_DMGR_PROFILE_HOME>/config/cells/<cell>/nodes/<node>/serverindex.xml
22 Compare the “serveridex.xml” files of the Deployment manager and local WAS server node for port conflicts under <WAS_DMGR_PROFILE_HOME>/config/cells/<cell>/nodes.
23 If port conflict occurs please change the ports for was server node in the <WAS_DMGR_PROFILE_HOME>/config/cells/<cell>/nodes/<node>/serverindex.xml
Remark: Keep common know port of deployment manager (9060,8879,9043) as it is.

24 Source DMGR Machine.
25 Stop the source DMGR node.
26 Remove the execute permissions for startmanager.sh script.

Target DMGR Machine.
27 SU to user who is required to start the dmgr like “was”
28 Start the new DMGR server with “startmanager.sh”.
29 check for logs if problems found in start.

Server Nodes :

30 Wsadmin utility in every node of the Cell should connect to ND; make sure ND host is specified for com.ibm.ws.scripting.host in the wsadmin.properties file.
31 <WAS_NODE_PROFILE_HOME>/properties/wsadmin.properties
32 Delete the content of the following sub-directories entirely.
33 <WAS_NODE_PROFILE_HOME>/wstemp
34 <WAS_NODE_PROFILE_HOME>/config/temp
35 After all documents for the node have been updated in DMGR profile configuration, execute the syncNode tool on each node in the cell containing the node being modified. SyncNode will replicate the configuration containing the changed "hostName" property to the local node profile. Once "syncNode" has been executed, the node should be functional again within the cell, with a new "hostName" property.
36 <WAS_NODE_PROFILE_HOME>/bin/syncNode (.bat/.sh) <DMGR_host> <DMGR_soap_port> -username <name> -password <password>
37 Note:”syncNode” command may take substantial time based on the size of the Cell, so give enough time to complete.
38 Start node agent by following command and verify that node agent is visible in the ND
39 Administrative Console. <WAS_NODE_PROFILE_HOME>/bin/startNode (.bat/.sh)
40 Start each server on node profile by following command.
41 <WAS_NODE_PROFILE_HOME>/bin/startServer (.bat/.sh) <server>
42 Start the node agents on the other servers
43 Start all the clusters
44 The exercise is said to fine only once the all the desired application servers are up and running.

Back Up Plan:
45 Failure to achieve desired result as mentioned in step 27 the backup plan is as follows :

46 Stop the DMGR on target server.
47 Start the DMGR on source servers.
48 Delete the content of the following sub-directories entirely.
<WAS_NODE_PROFILE_HOME>/wstemp and <WAS_NODE_PROFILE_HOME>/config/temp
49 Synchronize all the nodes with the source DMGR with “syncnode” command.
50 Start node agents.
51 Start clusters.