Many times customers try squeezing in all the tasks listed below into a single maintenance window or in some cases they are forced to squeeze in more tasks than originally planned (i.e. a file system check). The checklist below is intended to help to better plan your maintenance windows to make the best use of time to ensure the primary task of upgrading the LogLogic appliance is successful and without the additional stress of wondering whether it will complete on time. This will be accomplished by identifying tasks that can be performed by one or more separate maintenance windows. Any issues that arise by virtue of performing those other tasks will therefore not affect the maintenance window for the upgrade proper.

Hardware failures - First and foremost is to schedule a separate maintenance window for rebooting the appliance as close to the date/time of the maintenance window as possible for the upgrade itself. This serves two purposes:
1. If your LMI instance is on physical hardware then rebooting can uncover impending hardware failure because boot up sequences are capable of exposing them more than regular operating conditions.
2. Your appliance could be due for a file system check. It's better to let that occur on your schedule rather than being surprised by it occurring in the middle of an upgrade. Starting in LMI 5.7.0 the file system check is disabled by default but upgraded systems still have it enabled. The time threshold is 180 days so if your system hasn't been rebooted in over 180 days then the next reboot will trigger the fsck.

Hotfixes - Check the LMI ConfigureUpgrade guide for the version to which you are upgrading to see which hotfixes are required to be installed prior to upgrading. Hotfixes require an application restart but not hardware reboot. Keep in mind that sometimes hotfixes are required on the Management Center side which will be applicable if you use MC for deploying upgrades to your LMI appliances.

LSP - Check the LMI Release Notes to verify the minimum LSP version for the LMI version to which you are upgrading. Schedule a maintenance window for performing the LSP upgrade as well as any hotfixes. Each LSP installation requires an application restart. You may need to perform multiple LSP upgrades before reaching the minimum version for a given LMI release if the LSP was not already being kept up-to-date.

Management Center - If you have Management Center in your environment then there are 2 items you can do in preparation for an LMI upgrade:
1. As mentioned above, check the LMI ConfigureUpgrade guide to confirm if there is any MC-related hotfixes that are required.
2. Upload the LMI MC artifact files (everything in the TIB_logXXX_x.y.z_fileupgrade.tar file) to the MC repo. The LMI upgrade package gets a little bigger in each release (as of 6.3.0 the .tar file is 2.9GB)  so the size of the LMI artifacts for MC continues getting bigger. If the connection to your MC server is slow then it's best to upload the artifact files prior to your maintenance window.

Backups - The LMI ConfigureUpgrade strongly recommends performing a full backup of your LMI appliance prior to the upgrade. If your appliance is not currently configured for performing backups then TIBCO highly recommends configuring them so that there is something to fall back to in the event of a worst case scenario during the upgrade. A config only backup is adequate if you can withstand losing any event data still stored on local disk. Again, this is only in a worst case scenario. It's very rare to need a backup but it's nice to have one when it's needed. If you are upgrading an EVA then you can also create a VM snapshot. Both snapshots and LMI backups are best performed as close to the upgrade as possible to avoid a gap in log data in the event the snapshot or* backup has to be restored. LMI backups can be performed with the application still operating normally and collecting event data so a maintenance window is unnecessary.

Health check - The healthcheck script can be executed any time prior to the upgrade; it does not have to be executed immediately prior to it. It's best to execute the script at least a day or two before the upgrade in case there are any issues that it flags. This will give at least some time to address those issues, especially if TIBCO Support has to be involved. The healthcheck script has been enhanced in LMI 6.3.0 to perform additional checks. See the LMI 6.3.0 ConfigureUpgrade Guide for details regarding all the checks it performs. The healthcheck script verifies LSP version so it should be one of the last pre-upgrade tasks to perform.

Stage the upgrade files - The .tar.bz2 and .sig files can be staged in the /loglogic/update directory anytime prior to the upgrade. Again, just like with the MC artifact staging, if you have a slow connection between your workstation and the LMI appliances it may be best to upload the bz2 and .sig files to the appliance well before the upgrade maintenance window just to make the best use of the maintenance window duration. Uploading the file to one appliance from your workstation then copying it to the other appliances from the first appliance instead of directly from your workstation each time can also expedite the upload process for environments that have slow connections.

Pre-upgrade tasks or other upgrade considerations - It's highly recommended to always review the respective LMI version's ConfigureUpgrade guide before upgrading because it may contain specific information needed for the given LMI version that were not necessary or applicable in previous versions.

Hopefully the checklist above will help provide a more efficient upgrade process and reduce the risk to exceeding your maintenance windows.

This article provides a compiled list of items you can use as a checklist for preparing a LogLogic LMI appliance for an LMI upgrade.