This repository has been archived by the owner on Jul 16, 2020. It is now read-only.
Calls to docker APIs cannot be cancelled #2
Assignees
Labels
Comments
tpepper
pushed a commit
that referenced
this issue
Apr 22, 2016
This is probably not a real issue today, but... Each of our sample workloads appears meant to request 80MB of persistent storage. Which is ignored anyway. There is a typo for two docker workload descriptions which sets the resource_id to #2 for the disk line instead of Resource #2 is cpu, so technically the template is saying those docker workloads should get 80 CPUs worth of resource allocation. That's not good. Now really we don't account CPUs or create cpusets (that I'm aware of anyway) yet for docker workloads. And it may be that having read the earlier line requesting 2 CPUs worth of resource, that the second request is ignored. So we're probably multiple times lucky. Either way, let's just make sure we don't rely on luck.
kaccardi
pushed a commit
that referenced
this issue
Aug 29, 2016
osprepare: Remove redundant reslice of packages slice
|
Here's some background information about the context object |
Sign up for free
to subscribe to this conversation on GitHub.
Already have an account?
Sign in.
The reason is that when we call these functions we simply pass in a background context. The result is that the instance go routines that call these functions block until they return. This is only really a problem when shutting down launcher. A blocking call to a docker API may prevent ciao-launcher from shutting down properly when it receives a SIGTERM.
The text was updated successfully, but these errors were encountered: