Hi all,
Just as a safety net I'd like to publish copies of the databases powering my MediaWiki installations, so that if I lose all my data and backups, I might still be able to get something back (as well as reassuring my users that their contributions won't be lost forever!)
I don't want to publish the raw database dump because it contains people's e-mail addresses and password hashes, but I'm thinking that perhaps if I take a copy of the database and erase the values in those tables, I might be able to export and publish the 'anonymised' copy. I'd rather not omit the user table entirely, because if I do ever need to restore the wiki from this copy I'll need it to associate the right account with each edit.
Does anyone know which fields contain sensitive data that I should remove? I can see these obvious candidates:
user.user_password user.user_newpassword user.user_email user.user_token user.user_email_token
Are there any others that are potentially confidential and should not be made public?
Many thanks, Adam.
I can't think of anything offhand. You might want to run a database search to make sure none of those values in those user fields is available elsewhere on the database in cleartext, but otherwise, you should be fine.
To: mediawiki-l@lists.wikimedia.org From: a.nielsen@shikadi.net Date: Sat, 28 Sep 2013 22:51:38 +1000 Subject: [MediaWiki-l] How to anonymise MW database?
Hi all,
Just as a safety net I'd like to publish copies of the databases powering my MediaWiki installations, so that if I lose all my data and backups, I might still be able to get something back (as well as reassuring my users that their contributions won't be lost forever!)
I don't want to publish the raw database dump because it contains people's e-mail addresses and password hashes, but I'm thinking that perhaps if I take a copy of the database and erase the values in those tables, I might be able to export and publish the 'anonymised' copy. I'd rather not omit the user table entirely, because if I do ever need to restore the wiki from this copy I'll need it to associate the right account with each edit.
Does anyone know which fields contain sensitive data that I should remove? I can see these obvious candidates:
user.user_password user.user_newpassword user.user_email user.user_token user.user_email_token
Are there any others that are potentially confidential and should not be made public?
Many thanks, Adam.
MediaWiki-l mailing list MediaWiki-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/mediawiki-l
I can't think of anything offhand. You might want to run a database search to make sure none of those values in those user fields is available elsewhere on the database in cleartext, but otherwise, you should be fine.
Great! Thanks for the quick reply!
Cheers, Adam.
I suggest you ask Coren for this. He's in charge of the database of WMF Labs.
On Sep 28, 2013, at 8:51 PM, Adam Nielsen a.nielsen@shikadi.net wrote:
Hi all,
Just as a safety net I'd like to publish copies of the databases powering my MediaWiki installations, so that if I lose all my data and backups, I might still be able to get something back (as well as reassuring my users that their contributions won't be lost forever!)
I don't want to publish the raw database dump because it contains people's e-mail addresses and password hashes, but I'm thinking that perhaps if I take a copy of the database and erase the values in those tables, I might be able to export and publish the 'anonymised' copy. I'd rather not omit the user table entirely, because if I do ever need to restore the wiki from this copy I'll need it to associate the right account with each edit.
Does anyone know which fields contain sensitive data that I should remove? I can see these obvious candidates:
user.user_password user.user_newpassword user.user_email user.user_token user.user_email_token
Are there any others that are potentially confidential and should not be made public?
Many thanks, Adam.
MediaWiki-l mailing list MediaWiki-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/mediawiki-l
mediawiki-l@lists.wikimedia.org