Skip to content

Mitaka 21st march base latency optimized#416

Open
mak-454 wants to merge 17 commits into
mitaka_21st_march_basefrom
mitaka_21st_march_base_latency_optimized
Open

Mitaka 21st march base latency optimized#416
mak-454 wants to merge 17 commits into
mitaka_21st_march_basefrom
mitaka_21st_march_base_latency_optimized

Conversation

@mak-454

@mak-454 mak-454 commented Jun 3, 2016

Copy link
Copy Markdown

No description provided.

context, network_function_id)

self.active_threads.append(gth)
self.thread_list.append(th)

Copy link
Copy Markdown

Choose a reason for hiding this comment

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

Where is this "th" initialized? Also, where is this thread_list used?

Copy link
Copy Markdown
Author

Choose a reason for hiding this comment

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

cleaned up now.. this variables are not required.

result = getattr(driver, method)(**resource_data)
'''
# [AKASH]
result = const.SUCCESS

Copy link
Copy Markdown

Choose a reason for hiding this comment

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

Can be removed. Commented code.

Copy link
Copy Markdown
Author

Choose a reason for hiding this comment

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

done.

self.sc.poll_event_done(ev)
# self.sc.poll_event_done(ev)
self.notify._notification(notification_data)
return STOP_POLLING

Copy link
Copy Markdown

Choose a reason for hiding this comment

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

Didn't get the difference between returning STOP_POLLING vs invoking poll_event_done. Is this functionality fix or perf?

Copy link
Copy Markdown
Author

Choose a reason for hiding this comment

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

long time back "poll_event_done" was disabled, orchestrator moved to STOP_POLLING but configurator didnt. this change is not for perf.

@njagadish

Copy link
Copy Markdown

@ashutosh-mishra and @rajendramachani - can you please review your respective modules in this PR. These are perf related changes which needs to be merged soon. Please review ASAP.

def send(self, data):
self._socket.send(data)
self._socket.setblocking(1)
self._socket.sendall(data)

Copy link
Copy Markdown

Choose a reason for hiding this comment

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

Is this logic because we want to make send blocking but recv non-blocking? If so, why?

Copy link
Copy Markdown
Author

Choose a reason for hiding this comment

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

this was just a hunch while going over the components w.r.t perf activity that it could happen that proxy doesnt send up all the data to over the cloud.. need to verify this separately.

port)
return pt['port_id']

def get_neutron_port_details(self, token, port_id):

Copy link
Copy Markdown

Choose a reason for hiding this comment

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

This method name can be get_port_and_subnet_details. So that both nautron and gbp workflow call same method. It just looks like,

network_handler.get_port_and_subnet_details(token, port_id)

Is this port_id a neutron port ID ?. We should pass GBP pt id here in GBP workflow.

And in this method, get neutron port_id like,

_port_id = self.get_port_id(token, port_id)

Ref: get_port_details method in the same file.

Copy link
Copy Markdown
Author

Choose a reason for hiding this comment

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

RPM that is the optimization, we cannot do like that, please see "self.get_port_id" does "get_policy_target" which is an unnecessary REST API, while the caller has neutron port_id already.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

5 participants