WI07 Security Vulnerability handling

The purpose of this page is to provide instructions to the EGI Operations team members on how Security Issues handling should be done.

Previously this was managed by the IRTF team.

• For the time being, IRTF will still be responsible for looking at Pakiti/the security dashboard, look for false positives and create new tickets with a due

date of 3 days.

• If there is no acknowledgement or answer:
  • 1 working day before the due date, send another reminder via RT
  • .5 working day before the due date, send a last reminder, potentially including operational contacts in addition of security contacts. In such case, in case of answer, make sure to verify that the security contact is still valid
  • After the due date, suspend the site

• If there is an acknowledgement, but no solution announced:
  • 1 working day before the due date, check Pakiti:
    • If there is no change, ask for progress/for an update, insisting
    • If the vulnerability disappeared from Pakiti, ask for a confirmation that the vulnerability was fixed and that it's not simply the affected node not

being reached by the Nagios probe (which usually reach different nodes every day)

• After the due date, if there is still no answer/solution announced, suspend the site

• If a solution is said to be deployed:
  • If it's a mitigation:
    • If different from any mentioned in the advisory escalate to IRTF
    • If it's the same, wait for up to a day and check the security dashboard
  • If it's a simple package/kernel update, check pakiti:
    • If there is a report for the affected node(s) without any vulnerability, thanks and close the ticket
    • If the last report for the affect node(s) is still from before the update, ask to run the pakiti client by following https://wiki.egi.eu/wiki/EGI_CSIRT:Pakiti_client.
      • If the vulnerability then disappear from Pakiti, with or without any

other message, close the ticket

• If a solution is announced but for after the deadline, it can be tolerate if justified (e.g. special kernel update requiring different vendor update). In case of doubt don't hesitate to escalate to IRTF.

The main idea behind such handling is to make sure that sites are aware of the issue and working on it. We usually don't penalize sites that are showing good intention even if the progress is not within the procedure: SEC03 put deadlines on site to do few things, but does not put deadline on the CSIRT to enforce them, allowing us to adapt our answer.

I'm saying multiple time "escalate to IRTF", but as I understand there is no such system in RT, so pinging us by sending a mail to irtf@ should be enough.