openstack/nova
Code Issues Proposed changes

Remove stale aggregates notes from scheduler evolution doc

Browse Source 
Since I901184cb1a4b6eb0d6fa6363bc6ffbcaa0c9d21d in Kilo the
aggregates information about a HostState object (which is a
wrapper over a ComputeNode) is cached in the scheduler, so the
comments in the scheduler evolution doc about not accessing the
aggregates table in the DB from filters/weighers and such is
extremely out of date and should just be removed.

Change-Id: Ibcbad227813d3b37b4e314eddbf3bae6e85652ea
Related-Bug: #1820283
This commit is contained in:
Matt Riedemann 2019-03-15 11:39:33 -04:00
parent 0a44d3ae0a
commit 18c40cacc1
1 changed files with 0 additions and 15 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

15
doc/source/reference/scheduler-evolution.rst
View File

@ -43,14 +43,6 @@ that is close to the shared storage where that volume is. Similarly, for
the sake of performance, it can be desirable to use a compute node that the sake of performance, it can be desirable to use a compute node that
is in a particular location in relation to a pre-created port. is in a particular location in relation to a pre-created port.
Accessing Aggregates in Filters and Weights
--------------------------------------------
Any DB access in a filter or weight slows down the scheduler. Until the
end of kilo, there was no way to deal with the scheduler accessing
information about aggregates without querying the DB in every call to
host_passes() in a filter.
Filter Scheduler Alternatives Filter Scheduler Alternatives
------------------------------ ------------------------------
@ -91,10 +83,6 @@ pluggable through the service group API, and should be independent of the
scheduler service. For example, it could be managed via zookeeper rather scheduler service. For example, it could be managed via zookeeper rather
than polling the nova DB. than polling the nova DB.
There are also places where filters and weights call into the nova DB to
find out information about aggregates. This needs to be sent to the
scheduler, rather than reading directly from the nova database.
Versioning Scheduler Placement Interfaces Versioning Scheduler Placement Interfaces
------------------------------------------ ------------------------------------------
@ -124,9 +112,6 @@ This is linked to the work on the resource tracker.
Updating the Scheduler about other data Updating the Scheduler about other data
---------------------------------------- ----------------------------------------
For things like host aggregates, we need the scheduler to cache information
about those, and know when there are changes so it can update its cache.
Over time, its possible that we need to send cinder and neutron data, so Over time, its possible that we need to send cinder and neutron data, so
the scheduler can use that data to help pick a nova-compute host. the scheduler can use that data to help pick a nova-compute host.