The tool dumps your database and files into an Amazon S3 bucket created specifically for your migration. It runs directly on your Pure server, and there is no need to install it. It is a standardised tool that has been in use in Pure for many years to help with troubleshooting; you can find a version of it that copies the Pure database only in Pure at Administrator > Support > Database export. You will be sent a special version for hosting migrations (which copies both the database and files) that has been created for your locally hosted instance. On subsequent runs of the tool, it will skip any files that already exist (and have not been updated) in the S3 bucket. Files that are updated in Pure will have a new ID, and so be automatically captured and recopied.

Yes, the data is transferred to the S3 storage for your cloud server using a TLS-encrypted connection. Our encryption methods meet all security requirements of the AWS API endpoints (see AWS security blog post). The entire AWS operation is ISO27001 certified (see AWS services and certification).

You will run the tool at least twice: once at the beginning of your hosting migration project, to copy your locally hosted Pure into your new AWS server, and once at the end, to ensure the data in your AWS server is up to date when you make the final switch away from your locally hosted Pure.

It depends on the size of your Pure. For large Pures it may take more than a day. See also <Can I increase the speed at which the migration tool runs?>

You do not need to stop your Pure to run the tool. It will run in the background and does not noticeably affect performance. However, when you run the tool for the final time at the end of the hosting migration project, before you make the switch from your local server to the AWS one, you should stop your Pure server (meaning downtime for the backend) so that no changes are made to your data after it has been copied.

Before you run the tool, ensure you have an admin user set up to be authenticated with a normal Pure login (username and password), so that you can log in to your new Elsevier-hosted Pure. When your Pure is backed up to the cloud, external authentication methods may not work in the cloud Pure without further configuration. Alternatively, your Implementation Manager can set up a normal Pure login in your new Elsevier-hosted Pure and send you the password.

First, please check the Troubleshooting section below to see if there is an established solution to your error. If you cannot find one, contact your Implementation Manager, sending them a screenshot of the error and a log.

You can run the migration tool in ‘files only’ mode, meaning that only files will be copied, and not the database. Instructions are in the readMe in the migration tool package. The database is copied in full every time you run the tool and overwrites the version in your S3 storage, but if there are already files from a previous copy in your S3 storage, only new and updated files will be copied (files that are updated in Pure will have a new ID, and so be automatically captured). It is not possible to run the tool in a way that only copies changes to the database.

No.

Yes. The port is 443. You should receive a response from the Apache servers on the Pure IPs, which can be found in the Help Center at IPs in use for Pure and Pure Portal.

Yes. First check your Job Log Entries, as these can slow down the migration tool considerably – you can fix this by frequently running the "Prune Job Log Entries" cronjob. You can also increase the number of threads; instructions are in the readMe.

The exported data is kept in the S3 storage for your cloud server until two weeks after your migration, after which it is deleted permanently.

When you invoke the tool, ensure that the full command line forces -Dfile.encoding=UTF8 to be used as a Java VM option (not as an argument for the tool itself). This can be done by putting -Dfile.encoding=UTF8 right after the java command (java -Dfile.encoding=UTF8).

When running the migration tool for the first time, you can answer ‘y’ to this question without stopping your Pure: you are just creating a copy with which to test your new AWS server. There is a possibility of some broken data if your Pure is not stopped, since it will be writing to the database at the same time as the migration tool is reading it, but this is acceptable when you are testing. When you come to run the tool for the final time – before your Elsevier-hosted server goes live – you will need to stop your Pure in order to prevent broken data and also to ensure no further changes are made to your Pure after the copy is made. See < Should I stop my Pure/schedule downtime to run the migration tool?>.

Change the database connection string to "encrypt=true;trustServerCertificate=true".

• Is <Pure data directory> an expected value for the server data directory of Pure?
• If not, then the tool could have been pointed at a Pure database other than the one you intended to export.
• Was the export performed from the machine where the Tomcat process of Pure ran?
• If not, then that could explain why the tool was unable to find all files/directories.
• Is diagnostics present in the <Pure data directory>  directory?
• If not, then consider creating an empty the directory, such as by using mkdir <Pure data directory>/diagnostics (replacing <Pure data directory> in the command-line with the parent directory from the error message) and assigning appropriate permissions. This directory is not used when Pure is hosted.
• Was the export performed as an operating system user with the same privileges to read directories and files in <Pure data directory>?
• If not, then the export may be unable to access all directories/files in <Pure data directory>.

• Is <Pure data directory> an expected value for the server data directory of Pure?
• If not, then the tool could have been pointed at a Pure database other than the one intended to be exported.
• Was the export performed from the machine where the Tomcat process of Pure ran?
• If not, then that could explain why the tool was unable to find all files/directories.
• Is maintenance.html present in the <Pure data directory>  directory?
• If not, then consider creating an empty file, such as by using touch /data/pure_data/maintenance.html (replacing <Pure data directory> in the command-line with the parent directory from the error message) and assigning appropriate permissions. This file is not used when Pure is hosted.
• Was the export performed as an operating system user with the same privileges to read directories and files in <Pure data directory>?
• If not, then the export may be unable to access all directories/files in <Pure data directory>.