Attendees
RSV replacement status
- Derek has looked over the results
- Rates look good
- Data looks good
- There might be issues due to the loss of state when the container is restarted
- Does the first attempt fail, then the second one is more accurate?
- First collection needs to catch up on all the data
- What are the next steps to make this production?
- Shawn is ready to switch as long as the data is validated
- Monitoring
- Host “psrsv” on https://psetf.opensciencegrid.org/etf/check_mk/ (cert required)
- Query time: https://gracc.opensciencegrid.org/dashboard/db/perfsonar-collector?orgId=1
OSG check_mk
- Discussion on architecture
- Monitor with multiple servers? (using check_mk caching agent to avoid overhead)
- Use the distributed monitoring features?
- Shawn prefers to keep psetf focused on pS monitoring
Action items
- Derek: RSV replacement
- Map the state directory to the host to preserve across restarts (under /usr/local or /var)
- Run the new RSV collector for 24 hours and make a go/no-go decision on Mar 12, 2019
- John, along with Huijun/Derek: check_mk
- Work with Shawn on integration of OSG check_mk site with psetf
- Derek: Don’t send summary info to bus
- Shawn: Add pS ETF link onto sand-ci website
- Derek: Add collector status link to sand-ci website
- Shawn: Further polishing of the sand-ci OSG architecture document (PR #13)