Meeting Date

• TEAMS: 2024-10-02

Invited

Anthony (group leader), Lori, Clayton, Guoxiang, Nathan, Nick, Todd, Ed, Devon, Gwen, Ushe

Attendees: NL, AB, DM, NF, UU, EC, CT

Review and accept previous meeting minutes.

CsLWGMeeting20240918

Review last meeting's Action Items

linux.cscf.uwaterloo.ca

• Still blocked on comprehensive review of authentication/authorization configuration, work stalled - need to schedule in person meetings
  • Blocked on staff/manager availability

DFSc Performance

• Hourly snapshots enabled since Sept 19th.
  • Currently keeping 200+ hourly snapshots
  • Performance stable
  • Need to implement retention management, see RT: 1319807
  • Need to delete all old snapshots due to fossilized bad file ownership, see RT: 1319806
  • Documentation on WCMS and motd, see RT: 1321659, 1321659
• Should also email users who will have their homedirs deleted - Nick
  • RT: 1318097
• Quotas set for most accounts, list of quota violators provided for Nick, see RT: 1317978

Account Management

UID/GID re-use

• Currently causes problems with old homedirs winding up owned by new users, see RT: 1304698 and RT: 1319439
• Could fix, or we could just stop re-using UIDs/GIDs... Do we have any systems that don't support 32 bit UIDs/GIDs?
  • Reached out to Robyn Landers
    • Clayton has potentially fixed the problem of re-using UIDs/GIDs
    • Need to hammer out details what it looks like to not reuse UIDs/GIDs

New Business

• DFSc Migration
  • Will create tickets for individual clients as children of RT: 1321648
    • At a minimum will require adjusting the networking of each client machine, changing configuration of individual client services, and probably rebooting each client (hopefully) once
    • Not sure yet if or how much downtime there will be
    • Clients may have to be rebooted
    • Worst case scenario is whole DFSC needs to be taken down - would need to schedule this outside busy hours

New documentation about networking:

https://uwaterloo.atlassian.net/wiki/spaces/CSCF/pages/44269437014/Server+Room+Network+Refresh https://uwaterloo.atlassian.net/wiki/spaces/CSCF/pages/44269699151/Migration+Guide

At much later time of our choosing: 129.97.167.x needs to be moved (using upper half of IP space). Some services may want to build a new VM on newer OS anyway at that time.

Ubuntu 24.04 KVMs now available. No more LXCs, regardless of Ubuntu version.

We have physical servers (physical cloud hosts) in IST's server rooms, which we don't have access to. But Gwen might get Anthony, Devon, Guoxiang access to MC1061 (not sure about the EC2 one). If we can't get access, then might have to move our servers.

Comments