Joshua Hoblitt added a comment - 19/Sep/18 2:48 PM - edited summary of changes: the elcap esxi VMs were stopped, deleted, and removed from the jenkins agent configuration route53 / dhcp static reservations for the elcap VMs were removed the existing but broken sierra- [24] VMs were deleted the sierra- [13] VMs were each running a separate esxi node, so they were halted in turn to minimize disruption, and copied as a new VM rather than producing new/update to date sierra images as this should be done (instead) for high sierra in the near future. route53 / dhcp static reservations were updated/added for sierra- [24] sierra- [24] were added to the jenkins configuration all osx VMs had the memory increased from 16GiB -> 24GiB as there are now only 2 instances running per esxi node; 16GiB is in reserve for test instances the dhcp options were changed to us sqre-sw1.lsst.cloud as the ntp server rather than a list of noao ntp servers; this did not resolve the long running osx clock sync problem it turned out that neither esxi or sierra respects dhcp option 42; esxi hosts and all VMs and ntp manually [re] configured and forcibly sync'd the timezone on the osx VMs were set to America/Los_Angeles to be consistent with the ec2 linux VMs.