Linux Working Group

Meeting Date

• TEAMS: 2022-11-16

Invited

• Anthony (group leader), Clayton, Guoxiang, Lori, Fraser, Devon, Nathan, Nick, Todd, Dave, O

Attendees

• Anthony (group leader), Clayton, Guoxiang, Fraser, Devon, Nathan, Nick, Todd, O

Review and accept previous meeting minutes.

• CsLWGMeeting20221102

Review last meeting's Action Items

• suexec-flex https://rt.uwaterloo.ca/Ticket/Display.html?id=950947
  • Nathan to discuss with Lori and Issac to determine who should be rebuilding the package and redistributing the deb patch
    • Progress?
      • Dave's view of the long run, is that this work would be one of the responsibilities of the empty INF position. I just don't see current INF staffing being able to take this until later in Winter term (maybe as part of final Ubuntu 22.04 cleanup to go production).
    • Nathan gathering info from package authors and following up.
  • Anthony will interact further in ticket.
  • Need documentation for this project and some example test cases to motivate new developers.
• OpenVPN upgrade: containers and 2FA: https://rt.uwaterloo.ca/Ticket/Display.html?id=1231092
  • IST security triggered this.
  • It has happened and is running. Functionality is the same with the exception that it prompts for 2FA.
  • URL info in the RT above.
• Ubuntu 22.04 availability
  • No 100G ring access yet for LXC instances, waiting on layer 1 equipment
  • Dave will help Anthony with layer 1 needs.
    • This is now done.
• Separation of TEACHING and GENERAL VLANs: this is a CS legacy separation, do we believe IST's firewall config is sufficient to maintain this?
  • What exactly is the requirement here? Are we actually meeting the requirement at present? Have we ever?
  • O will add it to next cscf-dir, cscfmgm meeting agenda.
• Is there a landing page for first-time linux.student.cs.uwaterloo.ca ? O will work with erwarren on this.
  • Updates to be published on git.uwaterloo.ca/cscf/cscfcoop/... soon.
  • Progress? Ticket? O will populate.
• No snapshots on homedirs on linux.student.cs and there was an announcement sent out in August that they would be made available.
  • https://rt.uwaterloo.ca/Ticket/Display.html?id=1252069
  • There are clients asking about this.
  • The intent is to turn them back on and this may be delayed due to other priorities.
  • We need to respond to clients. O will work with cscfmgm to push a relevant general message out sooner if applicable.
  • O will bring this as an item to following cscfmgm meeting (they run hot and long already so there may be delays).

Power Outage Post Mortem [a2brenna, 30min]

• Thoughts?
• https://rt.uwaterloo.ca/Ticket/Display.html?id=1254651
• Mostly smooth operation in collaboration with MFCF as we share the Math server room with them and cooperate with their technical teams to maintain it.
• Most services were running smoothly during this work and that demonstrates the robustness of back-end service architecture (e.g. linux.student.cs ran smoothly throughout), as well as redundancy in power grid hook ups. More details are documented in Maser ticket above (RT#1254651).
• There were some issues with contractors not hooking up things after maintenance work was completed.
• vault.cs was down during this outage and related work, and impacted barbara.daly's CS course needs. It's not clear why as there are no known vault.cs dependencies to resources in M3. It was resolved eventually (https://rt.uwaterloo.ca/Ticket/Display.html?id=1255014). There are plans to improve vault.cs to prevent this in the future. * There may have been an impact to vault.cs due to its database back-end fail-over. It may be good to run some "fire drills", i.e., database failovers, between terms to generate/assess application failures. * Anthony, Guoxiang can assist with coordinating the drills with database sysadmins such as during December at some point. * Drills are great. * Can all db maintainers (postgres, mysql) meet up about this ? Possibly. * It appears that plant ops is initiating such outages more frequently than before.

Network simulation test question [O, if time available or next meeting]

• Context: https://uwaterloo.atlassian.net/jira/servicedesk/projects/IST/queues/issue/IST-72297
• Can I stress test a UW network router in a classroom with multiple persistent ssh connections from the same account, i.e. (while(N=1000) ssh linux.student.cs) ?
• Other stress tests ?
• Can I get the "stress" command available on linux.student.cs and only as a root or privileged group command ? see https://www.hecticgeek.com/stress-test-your-ubuntu-computer-with-stress/
• Started writing something here: https://git.uwaterloo.ca/cscf/wifi
• A stress test would also need to account for radio-interference between devices within the class.
• Devon's work with mac wifi provisioning may have overlap with this effort.

Comments