Skip to content

infosys GDPR compliance, BUGZ-3980

According to GDPR regulations we should avoid people's DNs in any record. The proposed fix is to use a non invertible hash.

This requires fixes in the:

These below are additional suggestions to ease the sysadmin and debugging tasks. They are not required for the patch to work, but may be a nice aid for debugging.

  • server side logs or controldir files if we want to map DN to hash in the logs @aleksandr.konstantinov status:?
    Example: (arex.log) 
       [2020-03-27 14:34:38] [Arc] [VERBOSE] [22276/3] Current jobs in system (PREPARING to FINISHING) per-DN (1 entries)
   [2020-03-27 14:34:38] [Arc] [VERBOSE] [22276/3] /DC=org/DC=nordugrid/DC=ARC/O=TestCA/CN=testexample (sha512:d760683d4227d31f388c7eaec87fe9ad3819f2416c614257f11afdaefc37ca27f14b6881ecc38258af66a41ca3002ac1594c48c4ca9ea5a9a804d4ff17d77d64): 1
    Example: (gridftpd.log) 
       [2021-02-02 17:23:21] [Arc.GridFTP_Commands] [VERBOSE] [13090/9416752] response: 220 Server ready\\
   [2021-02-02 17:23:22] [Arc.GridFTP_Commands] [INFO] [13090/139954999238336] User subject: /DC=org/DC=nordugrid/DC=ARC/O=TestCA/CN=testexample (sha512:d760683d4227d31f388c7eaec87fe9ad3819f2416c614257f11afdaefc37ca27f14b6881ecc38258af66a41ca3002ac1594c48c4ca9ea5a9a804d4ff17d77d64)
   [2021-02-02 17:23:22] [Arc.GridFTP_Commands] [INFO] [13090/139954999238336] Encrypted: true
    Example: (job.*.local, add a new field subjecthash) 
       [root@arc-iridium arc]# cat /home/arccontrol/job.wMtMDmlBtMynVALY5mGJwcyoABFKDmABFKDmNxGKDmABFKDmTahA1n.local
   [...]
   subject=/DC=org/DC=terena/DC=tcs/C=SE/O=Lunds universitet/CN=Florido Paganelli hep-frp@lu.se   
   subjecthash=sha512:d760683d4227d31f388c7eaec87fe9ad3819f2416c614257f11afdaefc37ca27f14b6881ecc38258af66a41ca3002ac1594c48c4ca9ea5a9a804d4ff17d77d64)
   [...]
  • arcctl if we want to provide ways to see these hashes @andrii status:?
    Example: 
       root@arc-iridium arc]# arcctl job info wMtMDmlBtMynVALY5mGJwcyoABFKDmABFKDmNxGKDmABFKDmTahA1n
        Name            : testsimple
        Owner           : /DC=org/DC=nordugrid/DC=ARC/O=TestCA/CN=testexample (sha512:d760683d4227d31f388c7eaec87fe9ad3819f2416c614257f11afdaefc37ca27f14b6881ecc38258af66a41ca3002ac1594c48c4ca9ea5a9a804d4ff17d77d64)
        State           : DELETED
        LRMS ID         : 5261
        Modified        : 2021-01-29 12:34:01

in CC for reference: @Konya @maikenp @waananen @ellert

!1122 (merged)

infosys

UPDATE: After a discussion on skype, I decided to implement just a minimal change where only nordugrid schema contains hashed DNs based on coreutils sha512sum. The resulting string is hex encoded.

example:

Subject: /DC=org/DC=nordugrid/DC=ARC/O=TestCA/CN=testexample
sha512sum: d760683d4227d31f388c7eaec87fe9ad3819f2416c614257f11afdaefc37ca27f14b6881ecc38258af66a41ca3002ac1594c48c4ca9ea5a9a804d4ff17d77d64

Another decision is whether to hash both nordigrid-arc-globalowner and GLUE2Owner, I would prefer to hash both to have consistency, but it has been currently not accepted by the rest of the team.

Edited by Florido Paganelli