common: Add a lame-duck mode to common/trace.c #8108
Draft
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
The tracing subsystem has some rather strong consistency requirements that are easily broken by incorrect use of its external API. For example not closing a span can cause an assertion on the next call to crash the daemon. Due to the separation in time and code, it is not always easy to debug there, and crashing the entire system is bad. The lame-duck mode is switched to as soon as the first error occurs, and we just stop emitting traces in that case. Changelog-None
cdecker
force-pushed
the
202507-trace-lame-duck
branch
from
5e5b582 to
3cedfaa
Compare
| @@ -532,6 +533,10 @@ def prinErrlog(node): | |||
| return 1 if errors else 0 | |||
|
|
|||
|
|
|||
| def get_trace_assert_errors(node): | |||
There was a problem hiding this comment.
Does this work for errors in stderr?
Suspect it might be better to have a function callback for tracing, so we can BROKEN log.
| */ | ||
| static bool active = true; | ||
|
|
||
| #define TRACE_ASSERT_RET(cond, msg, ret) \ |
There was a problem hiding this comment.
Macros that return are an anti pattern: just have a macro for trace_assert_fail which hands #cond, file and line to bool trace_assert_fail_ and open code it. That keeps it readable.
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Outdated suggestions cannot be applied.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.
The tracing subsystem is useful, but not useful enough to warrent
taking the entire daemon down when one of its internal consistency
requirements arent met. This commit introduces a lame-duck mode
that the node switches to as soon as it notices an inconsistency
and it prints the line and reason for the switch.
We can then later come back and investigate. The important thing
is that while the tracing instrumentation is incomplete, and can
cause these issues, we prefer no longer emitting traces rather
than taking the node down.