Migrating from VISULOX 3.x to 4.x | VISULOX Documentation

The following steps can be used to migrate from VISULOX 3.x to VISULOX 4.x .

In case of a redundant environment with more servers, that are updated in several steps,
make sure not to have servers running in the cluster with different VISULOX versions at the same time,

Preparation: Exporting data on the old servers

Export the database (use database clean command before), the cluster key and scripts in the tools directory.
Copy these files to another directory outside of VISULOX (/tmp/update is possibly too small). It is a good idea to use one folder for each server for the backup files.
```
mkdir /tmp/update
cp /opt/visulox/data/database/database.db /tmp/update
cp /opt/visulox/etc/key /tmp/update
cp /opt/visulox/etc/key.node /tmp/update
```
Backup or providing the files in the VISULOX Filestore
Backup customized files, e.g. event.sh, filecheck.sh, customized reports, pam.d, etc

Installation on new servers / running servers in parallel

Install new nodes with VISULOX 4.x and complete the setup (build cluster / standalone database and configuration)
Migration can be done online (makes sense for the audit data)

Copy the backup files to the node:

chown -R vlx: /tmp/update/

Dry run / check:
/opt/visulox/setup/update/migrate.tcl all -source /tmp/update/ -nodemap <old logical nodename>=<new logical nodename>

Migrate / run:
/opt/visulox/setup/update/migrate.tcl all -source /tmp/update/ -nodemap <old logical nodename>=<new logical nodename> -run

Getting the files and copy them to the new filestore with rsync (be aware of enough diskspace)
Migration detects if the database belongs to an Archive Server
Check / adjust previous saved customized files (event.sh, filecheck.sh, pam.d, etc)

migrate.tcl

The script executes a 1on1 migration of the complete audit data from all nodes
With the parameter -nodemap a mapping of the nodea name from OLD=NEW can be done
The switch allows to use either a mapstring or a file, that contains lines with a mapstring
OLD and NEW must be a logical nodename

Update same VISULOX Node

umount store and remove VISULOX 3.x
Install VISULOX 4.x and complete the setup (build cluster / standalone database and configuration)

Getting the backup to the node:

chown vlx: /tmp/update/
/opt/visulox/setup/update/migrate.tcl all -source /tmp/update/
/opt/visulox/setup/update/migrate.tcl all -source /tmp/update/ -run

migrate.tcl

The script executes a 1on1 migration of the complete audit data from all nodes

With the parameter -nodemap a mapping of the nodea name from OLD=NEW can be done
The switch allows to use either a mapstring or a file, that contains lines with a mapstring
OLD and NEW can be a hostname or a nodename
For example -nodemap vAAAA=vNEW1,vBBBB=vNEW2,server1.dom=server-new1.dom

Getting the files and copy them to the new filestore with rsync (be aware of enough diskspace)
Attach slot in store
Check / adjust previous saved customized files (event.sh, filecheck.sh, etc)

Finishing work

Check for errors / warnings

visulox-gateway status
visulox-portal status
visulox status
 
visulox integrity

VISULOX PORTAL Console

Adjust the Workspace application icons
Remove TransitWeb application from Workspace (Transfer from/to the client is now done in the File Transit Area in Workspace)

VISULOX Cockpit

Assign a category to action scripts manually, where the category is missing and assign these scripts again to the according policies
Check configured VISULOX Policies, Command Connetcs, Hosts, etc