Data Integrity

Hi All,

I am currently working on a validation project and have the following issue.

The application was fully validated on the validation server, before cut over to the production server.

On cutover to the production server the .ini file that contains the pathway to the correct database was performed incorrectly and was changed to the testing server which is not validated at all.

For the last 3 weeks all prodcution data is now being stored in the testing server which is not validated.

Now we want to transfer all of this data to the prodcution server.

Can this be done?

If the test server has not bring validated how can we ensure the integrity of the data.

Can we retro validate the test environment and then migrate the data?

Or will this data have to be disguard as we cannot ensure it’s integrity.

I would like to find out what the options are, as I am not sure myself.


Hmmm…thought I responded to this earlier…I would move the data. Do a deviation, incident report, or whatever you call it at your plant when things don’t go right. Then as part of the investigation, you can discuss data integrity and why it was not affected. If you guys didn’t know where the data was going, how could anybody else. Plus you can discuss what was occuring in the test environment, who had access, and accessed data. you probably need to discuss the aspects of the test server that are similar as well as disimilar from the qualified servers. I think that should cover it. Above else, document the issue and then what you did to correct and prevent. Oh wait that is CAPA. Yes, I think that does apply here.

Thanks Meyert

Sounds good, generating a Quality report as we speak to resolve this issue.


Hi all
I’m agreed with meyert to do a deviation. I work in a validation project that we transfer a data from a test server to a production server. We performed a statistical assessment to transfer all data from the test server (not validated) to the production server. In order to evaluate the reliability of the effectiveness and reliability of the data migration a sampling plan was performed using a Beta distribution. This assessment gave me the security that the all data migrated is corrected (in my case, based on the minimum reliability of 90% and confidence level of 95%). This required the verification of all the fields within each record taken at random throughout all the data migrated.

You might want to purge the data if you have ‘test data’ which is not part of the process data. A simple data migration plan can help you on this, which you need anyway, but modify it to ensure non critical data is not transferred.
I worked on a project where this occurred, but only for a few hours!