Resolving upgrade issues

If you are using any version of BMC Discovery 10.1 or later, the preferred method for upgrading a Patch, Service Pack, or version is from the appliance’s UI. For more information, see Upgrading BMC Discovery.

Upgrading from an earlier version

There is no upgrade path from BMC Discovery version 7.5 to version 10.1. Rather, you need to migrate your data to a new installation of BMC Discovery version 9.0 and then upgrade to BMC Discovery version 10.1.

If you are upgrading from an earlier 7.X or 8.X version you will first need to upgrade to one of the versions listed above.

Resolving upgrade issues

You must be logged in to the appliance using a screen session. Enter:

[tideway@localhost ~]$ screen 

See Recovering from a lost connection using screen.

In BMC Discovery 10.0 and later versions, you can perform the following upgrades on standalone machines and clusters either from the BMC Discovery UI or by running the tw_run_upgrade command line utility:
  • Upgrade to a BMC Discovery Service Pack or a version later than 10.0
  • Upgrade the Operating System

The preferred way to upgrade is through the UI instead of the tw_run_upgrade utility. The utility is recommended only if there is an upgrade issue and you are required to resolve it by using the options provided by the utility.

tw_run_upgrade capabilities

The tw_run_upgrade utility is an interactive command line tool. Based on the upgrade issue you have, it informs you about the option you must run the utility with and any additional action you are required to perform to resolve it. To learn about some of the typical scenarios where you will use the tw_run_upgrade utility, see Resolving an incomplete upgrade process.

Running tw_run_upgrade

Before you start an upgrade, make sure that you have downloaded the compressed upgrade archive from the BMC Electronic Product Distribution (EPD) site and copied that to the /usr/tideway/var/upgrade directory of the machine from which you will run the upgrade.

To run the upgrade using the tw_run_upgrade utility, you must login to the machine from which you will run the upgrade as the root user and type the following command:

tw_run_upgrade [options]

where options are any of the options described in tw_run_upgrade options and the common command line options described in Using command line utilities.

tw_run_upgrade options

The following table describes the tw_run_upgrade utility options which are not listed in Using command line utilities:

Command Line Option

Description

--auto

Specifies to suppress prompts for confirmation of continuing the upgrade after time estimation.

--distribute

Specifies to prepare the upgrade and distribute the upgrade files over the cluster. It does not run the upgrade process.
You can distribute the upgrade files first and run the upgrade later.

--fix-interrupted

Specifies to unlock the machine after a previous upgrade run was interrupted.

--no-version-check

Specifies not to check the upgrade compatibility.

--restart

Specifies to unlock the machine where the upgrade was interrupted and restart the upgrade on that machine.

--start

Specifies to prepare the upgrade, distribute the upgrade files (if not already distributed), and start the upgrade process. This is the default mode of execution.

--tmpdir=TMP_DIR_PATH

Specifies to set up the upgrade temporary directory.
If a machine is likely to run out of disk space, you can set a symlink for the upgrade directory to a different mounted file system which has free disk space.

Resolving an incomplete upgrade process

The following section contains user examples of some typical scenarios where you will use the tw_run_upgrade utility in a standalone machine and a cluster:

User example for a standalone machine

If the upgrade process is interrupted on a standalone machine, the command line directs you to run the upgrade process again by running the tw_run_upgrade --start command.

User examples for a cluster

  • Using the restart command option: While running the upgrade, if the upgrade process is interrupted on one of the machines in the cluster after the services were stopped, it puts that machine into a locked state and prevents the other members of the cluster (where the upgrade has completed) from starting. If you attempt to start the services manually from the command line of the machine where the upgrade was interrupted, it directs you to unlock it and resume the upgrade process by running the following command:

    sudo tw_run_upgrade --restart

    Running the command unlocks that machine and resumes the upgrade process. Once the machine is upgraded and it reboots, all the machines in the cluster will also start.

    Warning

    The upgrade must be restarted (tw_run_upgrade --restart) only if it fails at a stage during the actual execution of the upgrade script. If the upgrade fails at an earlier stage (for example, during preparing the upgrade or distributing the upgrade files on the cluster), restarting will only upgrade the machine from where you are running the upgrade.

  • Using the --fix-interrupted command option: While running the upgrade, if the upgrade process is interrupted on one of the machines in the cluster before the services were stopped, it puts that machine into a locked state and displays an error message on the machine from where you are running the upgrade:
    Member no longer aware of ADDM Upgrade operation
    The interruption stops the services only on the machine where the upgrade was interrupted. If you attempt to start the services on that machine manually from the command line, it directs you to unlock it by running the following command:

    sudo tw_run_upgrade --fix-interrupted

    Running the command unlocks that machine. The interactive command line tool informs you if any additional intervention is required before you can run the --start option and run the upgrade again for that machine.

Post upgrade steps

See post upgrade steps in Upgrading BMC Discovery.

As well as the notes on that page you should refer to the postupgrade_Vnn_TODO.log written out by the upgrade. This contains tailored advice of the tasks that must be completed on that particular appliance and these must be completed for correct future behavior.

Messages in the upgrade log

During the upgrade the firewall (iptables) is restarted. When a kernel upgrade is part of the upgrade, the firewall is unable to restart as there is a mismatch between the running kernel's version and the kernel on disk. The firewall logs a FATAL message, but as this is entirely expected, the upgrade wraps it in an information message: 
2011-07-25 09:36:46: INFO: FATAL: Could not load /lib/modules/2.6.18-53.1.14.el5 /modules.dep: No such file or directory 
This is expected behavior and does not indicate a problem with the upgrade.

Recovering from a lost connection using screen 

If you lose the connection to the appliance during the upgrade and you have used screen, you can reconnect to the appliance and recover the virtual terminal running the upgrade. To do this:

  1. Reconnect to the appliance and login as the tideway user.
  2. List the current screen sessions. Enter:

    [tideway@appliance01 ~]$ screen -ls
    There is a screen on:
    23274.pts-0.appliance01 (Detached)
    1 Socket in /var/run/screen/S-tideway.
    

    You can re-attach to it with a simple command:

    [tideway@appliance01 ~]$ screen -r

    The virtual terminal is recovered and you can see how the upgrade is progressing.

Was this page helpful? Yes No Submitting... Thank you

Comments