Linux Working Group

Meeting Date

• TEAMS: 2021-06-01

Invitees - Attendees

• Adrian, Devon, Fraser, Guoxiang, Lori, Nick, Omar, Todd, Anthony, Nathan

Review and accept previous meeting minutes.

Proposed Agenda Items

Monitoring and load management for CS246 Spring and Fall 2021 (Omar)

• Discussion with csk: 550 students in CS246 now, going up.
• Needs: mysql.cs, login servers
• Wants: local scratch disk on servers?
• To do: monitoring
• yc2lee: 20.04 student environment provisioned to run all necessary student tasks
• dmerner and a2brenna to finish Icinga install/config on linux.student.cs machines
• marmoset running on a container: monitor that
• Problems with marmoset go to yc2lee: Flowchart for student issue reporting: Course staff -> yc2lee -> INF
• a2brenna: end-to-end testing of workflow
• yc2lee and dmerner: develop and implement end-to-end testing for marmoset
• Single-client vs 1000-client issues: Difficult to simulate load
• cephfs 10^6 strays issue should be resolved in currently running Octopus version

July 17 power outage

• Get final decision on total shutdown out to users before the next meeting
• Suggest to users to complete tasks before July 17 to protect against date slippage
• 2-stage announcement: Initial now, 3 weeks before with definite plan

Cephfs snapshots

• When will they be available? Bring full hardware complement up first (unless pressing need)
• Snapshot schedule? vs backups
• Self-service file recovery is key
• Run parallel backup streams on different MDS regions of the filesystem
• cephfs fileystem can have 99 snapshots in it
• Period of snap regime: ~assignment?
• Used in CS246