Improves start up performance with lb templates #4379
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
|
hit #4381 |
tssurya
added
feature/services&endpoints
We were not waiting on for informers to fully shutdown for level-driven controllers. As a consequence, goroutines would still be running handling events while the next unit test case ran. In this case, a handler with no workqueue (like node_tracker) would attempt to read global config, while the next test case was modifying the global config, leading to a data race. Fixes: ovn-org#4381 Signed-off-by: Tim Rozet <[email protected]>
|
Rebasing on #4383 to see if that makes this unit test CI green |
Follow up from ovn-org#4270 During initial node sync, we were creating all of the lb templates over and over again in OVN. Service handler starts after, and will take care of adding missing LB templates while the node cache is already hot. There is no point in doing it during node info sync on start up. It is still required to do when nodes are added or removed post-start. Signed-off-by: Tim Rozet <[email protected]>
trozet
force-pushed
the
perf_improvement_lb_templates
branch
from
efcd6c4
to
f1f8ef1
Compare
| // Stop stops the factory informers, and waits for their handlers to stop | ||
| func (wf *WatchFactory) Stop() { | ||
| klog.Info("Stopping watch factory") | ||
| wf.iFactory.Shutdown() |
There was a problem hiding this comment.
note for others: for the shutdowns here, stop channels must be closed prior to calling shutdown, and that is accomplished before Stop() func being called.
| c.nodeIPv6Templates.AsTemplateMap(), | ||
| } | ||
| if err := svcCreateOrUpdateTemplateVar(c.nbClient, nodeIPTemplates); err != nil { | ||
| klog.Errorf("Could not sync node IP templates") |
There was a problem hiding this comment.
include the error message in the log
| if err := svcCreateOrUpdateTemplateVar(c.nbClient, nodeIPTemplates); err != nil { | ||
| klog.Errorf("Could not sync node IP templates") | ||
| return | ||
| c.startupDoneLock.RLock() |
There was a problem hiding this comment.
note for others - this will not block for long. Another controller only locks to write to a bool startupDone and unlocks immediately after.
| klog.Errorf("Could not sync node IP templates") | ||
| return | ||
| c.startupDoneLock.RLock() | ||
| defer c.startupDoneLock.RUnlock() |
There was a problem hiding this comment.
do we need to keep hold of the lock while processing svcCreateOrUpdateTemplateVar below?
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Follow up from #4270
During initial node sync, we were creating all of the lb templates over and over again in OVN. Service handler starts after, and will take care of adding missing LB templates while the node cache is already hot. There is no point in doing it during node info sync on start up.
It is still required to do when nodes are added or removed post-start.