-
Notifications
You must be signed in to change notification settings - Fork 4.3k
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
change default bench-tps
client to tpu-client
#35335
Conversation
bench-tps/src/cli.rs
Outdated
Arg::with_name("tpu_client") | ||
.long("use-tpu-client") | ||
Arg::with_name("thin_client") | ||
.long("use-thin-client") | ||
.conflicts_with("rpc_client") | ||
.takes_value(false) | ||
.help("Submit transactions with a TpuClient") |
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.
We should preserve the --use-tpu-client
flag (and the various conflicts_with
logic), even though it duplicates the (new) default behavior, to avoid breaking downstream users.
bench-tps/src/cli.rs
Outdated
Arg::with_name("thin_client") | ||
.long("use-thin-client") | ||
.conflicts_with("rpc_client") | ||
.takes_value(false) | ||
.help("Submit transactions with a TpuClient") | ||
.help("Submit transactions with a ThinClient") |
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.
Actually, if we're going to remove this imminently, I think we should avoid exposing a flag for it in the first place.
If there is an internal need to keep it for a bit (eg /net
scripts), let's mark this as .hidden(hidden_unless_forced())
so it's not visible. If there's no need for it, let's just remove this Arg.
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.
Ya if we remove it here we'd have to remove all references to ThinClient
in /net
scripts in this PR as well. So maybe as you said, let's use: .hidden(hidden_unless_forced())
for now. And then for the PR where we will deprecate ThinClient
, we can fully remove this flag and remove all references to ThinClient
in /net
. does that work?
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.
Yes, that works for me.
Codecov ReportAll modified and coverable lines are covered by tests ✅
Additional details and impacted files@@ Coverage Diff @@
## master #35335 +/- ##
=======================================
Coverage 81.7% 81.7%
=======================================
Files 834 834
Lines 224235 224235
=======================================
+ Hits 183390 183391 +1
+ Misses 40845 40844 -1 |
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.
r+ one last little nit
bench-tps/src/cli.rs
Outdated
.conflicts_with("tpu_client") | ||
.takes_value(false) | ||
.hidden(hidden_unless_forced()) | ||
.help("Submit transactions with a ThinClient") |
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.
Maybe add a note here that usage of ThinClient is discouraged and that ThinClient will be deprecated?
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.
lgtm - gave it a quick test to make sure perf (for my single node use case) was unaffected.
* change default bench-tps client to tpu-client * remote client default to tpu-client * add --use-tpu-client back in. hide --use-thin-client * address nit, inform of future thinclient deprecation
* change default bench-tps client to tpu-client * remote client default to tpu-client * add --use-tpu-client back in. hide --use-thin-client * address nit, inform of future thinclient deprecation
Problem
It is time to deprecate and then remove
ThinClient
frombench-tps
. This is the first PR to simply just change the defaultbench-tps
client fromThinClient
toTpuClient
Summary of Changes
Change
ThinClient
toTpuClient
, add flag for runningthin-client
explicitly (will be removed in future), and updatenet.sh
scripts to reflect the change.Fixes #