Commit Graph

17 Commits

Author SHA1 Message Date
Artyom Lukianov
73a5cce3e6 device manager: do not clean admitted pods from the state
Signed-off-by: Artyom Lukianov <alukiano@redhat.com>
2021-08-08 16:46:06 +03:00
Elana Hashman
ee0bcac1d2
Migrate devicemanager/topology_hints.go to structured logs 2021-03-15 12:39:45 -07:00
Byonggon Chun
9da0912a33 Implement devicemanager.GetPodLevelTopologyHints() function
* Add podDevices() func
* Add getPodDeviceRequest() func

Signed-off-by: Krzysztof Wiatrzyk <k.wiatrzyk@samsung.com>
2020-11-12 12:25:55 +01:00
Krzysztof Wiatrzyk
6db58b2e92 Update logging to use a format util
Signed-off-by: Krzysztof Wiatrzyk <k.wiatrzyk@samsung.com>
2020-11-12 12:25:55 +01:00
sw.han
f5997fe537 Add GetPodTopologyHints() interface to Topology/CPU/Device Manager
Signed-off-by: Krzysztof Wiatrzyk <k.wiatrzyk@samsung.com>
2020-11-12 12:25:54 +01:00
Kevin Klues
00df26a985 Fix a bug whereby reusable CPUs and devices were not being honored
Previously, it was possible for reusable CPUs and reusable devices (i.e.
those previously consumed by init containers) to not be reused by
subsequent init containers or app containers if the TopologyManager was
enabled. This would happen because hint generation for the
TopologyManager was not considering the reusable devices when it made
its hint calculation.

As such, it would sometimes:
1) Generate a hint for a differnent NUMA node, causing the CPUs and
devices to be allocated from that node instead of the one where the
reusable devices live; or
2) End up thinking there were not enough CPUs or devices to allocate and
throw a TopologyAffinity admission error

This patch fixes this by ensuring that reusable CPUs and devices are
considered as part of TopologyHint generation. This frunctionality is
difficult to unit test since it spans multiple components, but an e2e
test will be added in a subsequent patch to test this functionality.
2020-07-20 11:41:13 +00:00
Kevin Klues
74fe9364c3 Simplify logic in devicemanager TopologyHint generation 2020-07-20 11:41:13 +00:00
Davanum Srinivas
442a69c3bd
switch over k/k to use klog v2
Signed-off-by: Davanum Srinivas <davanum@gmail.com>
2020-05-16 07:54:27 -04:00
Takeaki Matsumoto
785fac6826 Make updateAllocatedDevices() as a public method and call it in
podresources api
2020-02-07 13:26:56 +09:00
Kevin Klues
bc686ea27b Update TopologyManager.GetTopologyHints() to take pointers
Previously, this function was taking full Pod and Container objects
unnecessarily. This commit updates this so that they will take pointers
instead.
2020-02-03 17:13:28 +00:00
Kevin Klues
4d4d4bdd61 Ensure devicemanager TopologyHints are regenerated after kubelet restart
This patch also includes test to make sure the newly added logic works
as expected.
2019-11-06 15:01:34 +00:00
Kevin Klues
a338c8f7fd Add some more comments to GetTopologyHints() in the devicemanager 2019-11-05 13:06:23 +00:00
Kevin Klues
58f3554ebe Sync all CPU and device state before generating TopologyHints for them
This ensures that we have the most up-to-date state when generating
topology hints for a container. Without this, it's possible that some
resources will be seen as allocated, when they are actually free.
2019-11-05 13:00:20 +00:00
Connor Doyle
e35301c19f Rename package socketmask to bitmask.
- As discussed in reviews and other public channels,
  this abstraction is used to represent numa nodes, not sockets.
- There is nothing inherently related to sockets in this package anyway.
2019-09-23 17:08:45 -07:00
Kevin Klues
eb0216e54e Update semantics to set Preferred field in TopologyHint generation
We now only set Preferred to true if resources can be allocated with a
size equal to the minimimum _possible_ mask when all resources are
available.
2019-08-29 14:32:10 -05:00
Kevin Klues
d3d7a8f5d4 Generate TopologyHints from the devicemanager 2019-08-29 07:45:43 -05:00
Louise Daly
9a118ceac4 Added stub support for Topology Manager to Device Manager
Co-authored-by: Conor Nolan <conor.nolan@intel.com>
Co-authored-by: Sreemanti Ghosh <sreemanti.ghosh@intel.com>
Co-authored-by: Kevin Klues <kklues@nvidia.com>
2019-08-29 07:45:43 -05:00