A backup that restores last night's BMS server proves nothing about opening six-year-old pressure records after an upgrade.
EU GMP Annex 11 handles the two in separate clauses, and they need different tests.
Clause 7.2 asks for regular backups of relevant data, with the integrity and accuracy of the backup and the ability to restore it checked during validation and monitored periodically. A restore test answers it: the server comes back, the application starts, last week's data is there.
Archiving sits in clause 17. Archived data has to be checked for accessibility, readability and integrity, and when the system changes, whether new hardware or a new software version, the ability to retrieve the data must be ensured and tested. Retention runs for years, so records are often read on a system that no longer looks like the one that wrote them.
The gap usually opens at an upgrade. The backup job is green every night, and nobody tries to open the old historian files with the new version until an investigation needs them.
One retrieval test worth adding to every upgrade: pick a trend from the oldest retained year, open it in the new version and compare a few values with the exported or printed record.
Our automation engineers design SCADA and BMS to 21 CFR Part 11 and Annex 11, with the URS and validation strategy written alongside the design.
When did your site last retrieve archived trend data after a system change?
Industries




