Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

Combine batches of successive roles for same nodes, #1345

Open
wants to merge 1 commit into
base: master
Choose a base branch
from
Open

Combine batches of successive roles for same nodes, #1345

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
32 changes: 32 additions & 0 deletions crowbar_framework/app/models/service_object.rb
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -919,6 +919,36 @@ def self.proposal_to_role(proposal, bc_name)
RoleObject.new role
end

# we can speed-up the application of (n+1)th role if both(n,n+1)
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Thanks for this comment block - definitely important to have it since this batch merging is not immediately obvious without explanation!

# roles are being applied on the same node
#
# eg. In our 2node deployment ceilometer{server,central} are always
# applied on the same node, given that they have different priorities,
# they are to be applied, one after the other, these priorities come
# from element_run_list_order
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Two nitpicks here:

  1. The priorities come from chef_order by default but can be overridden per role by element_run_list_order
  2. ceilometer-{server,central} is just one example, but this code will merge any adjacent batches regardless of what run_list priority the roles in. Is it possible to make the code automatically check that the run_list priority of batch n is lower than the priority of batch n+1, and only merge if that is true? If that was added then I think this would feel safe enough to me for a +1 vote.

#
# In other words: it's actually reducing the number of times chef-client
# is run rather than speeding up execution of any single run, by
# merging batches together
#
# a batch is [roles, nodes]
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Thanks - I find comments like this really helpful for understanding the data structures being manipulated.

def merge_batches(batches)
merged_batches = []
unless batches.empty?
current_batch = batches[0]
batches[1..-1].each do |next_batch|
if next_batch[1] == current_batch[1] && !current_batch[0].nil?
current_batch[0] << next_batch[0]
next
end
merged_batches << current_batch
current_batch = next_batch
end
merged_batches << current_batch
end
merged_batches
end

#
# After validation, this is where the role is applied to the system The old
# instance (if one exists) is compared with the new instance. roles are
Expand Down Expand Up @@ -1171,6 +1201,8 @@ def apply_role(role, inst, in_queue, bootstrap = false)

batches << [roles, nodes_in_batch] unless nodes_in_batch.empty?
end

batches = merge_batches(batches)
Rails.logger.debug "batches: #{batches.inspect}"

# Cache attributes that are useful later on
Expand Down