-
Notifications
You must be signed in to change notification settings - Fork 3.9k
Detect transport executors with no remaining threads #11503
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.
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
Merged
Merged
Changes from all commits
Commits
Show all changes
8 commits
Select commit
Hold shift + click to select a range
5ae7079
Detect transport executors with no remaining threads
shivaspeaks 1a56f71
OkHttp: Detect transport executors with no remaining threads
shivaspeaks a6201c5
OkHttp: Detect transport executors with no remaining threads
shivaspeaks 55f567c
Resolving blockers
shivaspeaks de44623
Resolving blockers
shivaspeaks 10fbc53
Detect transport executors with no remaining threads
shivaspeaks 2fd24b6
Detect transport executors with no remaining threads
shivaspeaks 4d44e3d
Detect transport executors with no remaining threads
shivaspeaks File filter
Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
There are no files selected for viewing
This file contains hidden or 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
This file contains hidden or 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
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Outdated suggestions cannot be applied.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.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I would like to understand the reason for this previous
latch
.as it seems, this code waits for
sendConnectionPrefaceAndSettings
execution, which happens below.Can that be moved above this
serializingExecutor.execute(new Runnable()
?Have I missed anything?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Writes are performed by AsyncSink. When a thread wants to do a write, it is added to a queue and a Runnable is added to
serializingExecutor
. So the Runnable here is running on that "same thread" and the writes can't happen until this proceeds.The question I had a bit earlier yesterday was, "why don't we do sendConnectionPrefaceAndSettings() in this runnable, instead of waiting on the latch?" This construction guarantees that the first things written after the TCP/TLS handshake is the HTTP/2 handshake. Back when this code was written, RPCs would be sent on transports before the transport went READY, so the sendConnectionPrefaceAndSettings() needed to be enqueued before start() returned.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
The introduction of DelayedClientTransport avoided using transports before they were READY. TransportSet is known as InternalSubchannel today. The activeTransportFuture.set() right after start() was the main case that sent RPCs to transports before they were ready.
cf787bd#diff-c993808318f59c8a206c5b4f1af4fd2b3f81a0fca3662be8966c392e3829e430R200
I feel like it took some years to be confident we had gotten rid of all the places that assumed a transport could support RPCs immediately after being started. But that was the most important change in that direction. The change was prompted by repeated memory leaks when dealing with ListenableFutures, as you can't remove listeners from futures. And in the process we fixed assigning RPCs too eagerly to transports.
(There had been some debate at the time whether we were okay with the added latency. Waiting for transport ready actually delays RPCs from being sent on a new connection by a RTT, because it waits to receive HTTP/2 SETTINGS from the server.)