coclobas
coclobas copied to clipboard
Configurable Cloudy Batch Scheduler
The line type nonrec exn = exn [@@deriving show] (https://github.com/hammerlab/coclobas/blob/master/src/lib/aws_batch_job.ml#L28) May fail with future versions of `ppx_deriving`.
Context: #57 Basically, the issue is that the new **default** `gci` image has a completely new way of managing the node by making use of a minimal distribution that lacks...
Of course, they reinvent variables within their JSON job-descriptions: http://docs.aws.amazon.com/batch/latest/userguide/job_definition_parameters.html#parameters ![image](https://cloud.githubusercontent.com/assets/617111/25635743/37ee027c-2f4d-11e7-984e-272ab7b959a1.png) ![image](https://cloud.githubusercontent.com/assets/617111/25635858/914743b0-2f4d-11e7-862c-e31c269d2d30.png)
We should keep track of the killing process until it's "acknowledged" by the backend.
I can view a job's logs soon after it has completed, but they usually disappear. ![kubectl-logs grab](https://cloud.githubusercontent.com/assets/6350854/24870230/a875daea-1de3-11e7-85c0-f6d8171e5a60.jpeg) Sometimes the job in question will have an empty dir in `/coclo/_cocloroot/logs/logs/job`. For...
e.g. to better be able to report this issue https://github.com/hammerlab/ketrew/issues/509#issuecomment-273293950
`kubectl` on Google cloud has so many ways of failing that going to find the logs has become too cumbersome.