• Nebula project allocations:
    • Sent request to nebula@ncsa.illinois.edu on Monday and spoke with Chris personally. This is still not done. We can't deploy the beta cluster without it.
  • NDS-404/NDS-406: Beta cluster deployment:
    • The longer we wait on this, the higher the risk.  Particularly since we're relying on the security group for wildcard cert. This is currently blocked by the Nebula project changes and NDS-602.
  • NDS-607: Docker cache using emptyDir
    • Testing the Tool Server (NDS-588) has revealed that our local Docker cache (NDS-343) is using an "emptyDir" config for storage.  This means that the cache has never been tested the Gluster storage.  I've logged NDS-607 for this. We can disable the cache but this will make the Tool Manager demo quite slow.  Also, this seems to be a repeatable test case for NDS-535.
  • NDS-535: "no data available" – corruption error 
    • Enabling Gluster storage for the Docker cache, the "no data available" error reported in NDS-535 is no longer intermittent. NDS-535 is a data corruption error in Gluster. We do not yet understand the root cause. It's surprising that testing NDS-262 hasn't revealed this.
  • NDS-602: /var/lib/docker volumes
    • David spoke with Chris on Monday and recommended that we modify the ansible provisioning scripts to add a local volume to each node (instead of relying on compute node storage).
  • NDS-579: Performance testing
    • As discussed earlier this week, we might consider prioritizing manual testing in the short term since automation is taking longer than anticipated
  • "Tool Manager Demo"
    • No ticket for this yet – but clearly this will take away from other system testing/hardening.
  • No labels