Applies To:Show Versions
- 1.1.1 PTF-02, 1.1.1 PTF-01, 1.1.1, 1.1.0
This appendix contains:
- Backup strategy
- Command line utility terminology
Backup and restoration of the GLOBAL-SITE Controller is currently performed using the command utility. This section provides procedures for backup and restoration under four potential situations.
We recommend that you back up only the configuration database (/gSITE/Configdb). Use the backup procedure that you prefer and are familiar with, and work with your F5 field service engineer to set it up.
We provide procedures for restoring the GLOBAL-SITE Controller in four types of situations:
- One repository (section) on the primary/publisher is corrupted.
- There is a complete system failure or database corruption on the primary (publisher).
- A single repository/section on a distributor is corrupted.
- There is a complete system failure or database corruption problem on a distributor.
One repository (section) on the primary (publisher) is corrupted.
- To clean out the repository/section, run this command on the primary/publisher to remove this directory and all subdirectories:
rm -rf /gSITE/Data/GSID<local IID>/<group name>/Repos/<section name>
- For each publication that uses this section, create an edition with version -1 for that repository/section (gsCreateEdition). All other sections must have their currently published version in this edition.
- Use gsSetPublished to set all paths to the new edition in each applicable publication.
- Delete all previous editions in each applicable publication, as they now refer to versions for this section that do not exist, but possibly will soon exist again with different content.
- Do a section update for that repository/section. This will recreate and repopulate the directory that was removed above.
If the stage is not in sync with the target/subscriber paths (perhaps untested and/or unpublished modifications have been made), then one of the target /subscriber paths must be used temporarily as the source path for this section. If you use a temporary source path, once the update is finished, remember to set the source path back to the original path.
- If there are any distributors, you must copy the directory mentioned in step 1 (and its subdirectories) to each distributor, putting the distributors into sync with the primary's/publisher's repository/section.
- For each applicable publication, create an edition with version 1 for that repository/section (just updated with version 1) and with the published version for all other sections.
- Set all paths as published (gsSetPublished) with this new edition in each applicable publication. (This step assumes that all paths for this section were published with this version of this section. If not, you must do an actual publish.)
The effect of this restoration is that all previous versions of this repository/section have been lost.
There is a complete system failure or a database corruption problem on the primary (publisher).
- Restore the configuration database.
This may require that /gSITE/Configdb be writable via NFS, and may require assistance from the field engineer.
- Update any changes made since the last backup of the database that are pertinent or needed immediately.
- Perform the following steps using the procedures from Scenario one:
a) Perform step 1 of Scenario one for each section.
b) Create an edition with version -1 for all sections in all publications.
c) Perform steps 3 and 4 of Scenario one.
d) Perform a section update for all sections using the instructions in step 5 of Scenario one.
e) Perform step 6 of Scenario one for each section.
f) For each publication, create an edition with version 1 for each applicable section (just updated with version 1).
g) The same as step 8 of Scenario one.
On a distributor, a single repository (section) is corrupted.
- On that distributor, to clean out the repository (section), run this command:
rm -rf /gSITE/Data/GSID<primary IID>/<group name>/Repos/<section name>
- Copy that same directory and all of its subdirectories from the primary/publisher to the distributor.
On a distributor, there is a complete system failure or database corruption problem.
- Restore the configuration database and/or rebuild by hand (via the command utility).
- For each section that this distributor distributes, go through the steps in Scenario three.
This section lists terms used primarily in the command utility.
For a complete list of terms used in the browser-based application, check the Glossary section.
Target servers for sets of files. The subscriber, where content is available to the client's customers.
An FTPaccess object name is a specific set (consisting of a node, user name, and password) used for FTP access to a node.
An FTPpath is a particular node, user name, and path. This is the directory path of either the source or target for publishing with the GLOBAL-SITE Controller.
The name of a source or target for an operation in the GLOBAL-SITE Controller publication process. The node can be the IP address or the host name.
Refers to one of the scheduled publishing options, used but not named, in the browser interface. Manual publishing is user-initiated, daily publishing happens at specified times during specified days of the week, polled publishing is continuous publishing, where the publishing process occurs at timed intervals, for instance, every 20 minutes, throughout the specified days of the week.
In the browser-based application, referred to as a publisher.
In the browser-based application, referred to as publication.
In the browser-based application, referred to as subscriber.
In the browser-based application, referred to as distributor.
In the browser-based application, referred to as section.
Source servers that house files that will be included in a site.
The server where the GLOBAL-SITE Controller delivers content to be viewed by the client's customers.