Github user joshelser commented on the pull request:
https://github.com/apache/accumulo/pull/94#issuecomment-210495111
> This should allow proper tracing propagation across the binning thread
pool, and consequently prevent loss of "sendMutation" events.
> In TabletServerBatchWriter.MutationWriter.queueMutations(...) (line 736),
HTrace Span's are not properly propagated across thread boundaries.
Consequently, tracing doesn't propagate into
TabletServerBatchWriter.SendTask.send(...), and so the sendMutation event can
fail to be logged.
So, IIRC, Tracer maintains the current Span via a ThreadLocal. So, I could
see spawning a new Thread in the binning process might lose that, and the
"lower" half of the span would be un-rooted (the part doing binning and
sending).
A couple of questions:
* Any thoughts on how we could make a test which would specifically check
for regressions here?
* Have you been able to notice your changes positively affecting
`ShellServerIT#trace`'s success rate? I have seen it fail now and again, but
never reliably.
* Any idea as to why we only *sometimes* see ShellServerIT fail? Is there a
reason why this doesn't always fail that you noticed?
Thanks!
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at [email protected] or file a JIRA ticket
with INFRA.
---
