The procedure how to upgrade CPUSE agent was totally replaced by the new one, while the old instructions were deleted.
In mentioned SK there were imstructions for older versions how to install it manually. Now, there is even no notice that it was replaced/removed/deleted...
You have to search google...
the "old" procedure was as follows:
Procedure:
- Transfer the CPUSE Agent package (DeploymentAgent_<build>.tgz) to the machine (into some directory – e.g., /some_path_to_CPUSE/).
- Unpack the CPUSE Agent package:
[Expert@HostName:0]# cd /some_path_to_CPUSE/
[Expert@HostName:0]# tar -zxvf DeploymentAgent_<build>.tgz
- Install the CPUSE Agent RPM:
[Expert@HostName:0]# rpm -Uhv –force CPda-00-00.i386.rpm
Notes:
- During the installation, you will see the « Waiting for DAService to stop… » message.
The preinstall script tries 60 times to stop the DAService executable (by issuing the « dbget installer:stop » command).
If the script does not succeed to stop the DAService, then it kills it and the installation continues.
- Related solution: sk119954 – « error: Failed dependencies » during manual installation of CPUSE Agent package
- [Optional] Restart all Clish daemons and ConfD daemon:
To improve CPUSE Agent, new CPUSE commands are added to Clish from time to time.
It is necessary to restart all Clish daemons for new CPUSE Clish commands to become available, and to restart ConfD daemon.
- If your default shell is CLISH (/etc/cli.sh), then log in to Expert mode.
If your default shell is BASH (/bin/bash), then log out from all Clish sessions.
- Kill all running Clish daemons (they will be restarted automatically):
[Expert@HostName:0]# killall -v clish clishd
Important Note: If your default shell is CLISH (/etc/cli.sh), then after issuing this command, you will be disconnected / logged out from all active sessions.
- Restart the ConfD daemon:
- [Expert@HostName:0]# tellpm process:confd
- [Expert@HostName:0]# tellpm process:confd t
Maintenance window is required:
- When confd daemon is starting, by design, it restarts any currently running routed daemons (by sending a TERM signal). It is done to avoid possible issues in Gaia Clish (e.g., returning invalid results for routing-related commands like «
show route
« ).
- Since routed daemon is responsible for all the routing in Gaia OS, short traffic outage will occur while routed daemon is being restarted.
- Since routed daemon is a Critical Device in Check Point cluster (since R76), cluster fail-over might occur while routed daemon is being restarted (refer to sk92878).
- Start the CPUSE Agent manually: [Expert@HostName:0]# $DADIR/bin/dastart
Kind regards,
Jozko Mrkvicka