[lttng-dev] lost trace events

[Jonathan Rajotte-Julien]

> A custom trace event class emits "begin" and "end" events in its
> constructor and destructor respectively. So I do not think this is due to
> conditional path.

If you can guarantee that the destructor is called, seems good to me.

Do you see scenario where the "begin" event is missing ?

> 
> Sequence of commands to capture a snapshot are:
> lttng create --snapshot
> lttng enable-event --userspace --all
> lttng add-context --userspace -t vpid -t vtid
> lttng start
> lttng stop
> lttng snapshot record

Nothing out of the ordinary here.

> 
> We create a new session for each snapshot to make sure that the buffer does
> not have any left-over events from the previous session. But I am curious
> if lttng shares buffers between sessions or re-uses buffers from an old
> session without flushing?

No sharing is done between sessions AFAIK.

> 
> Note that I am using the default LTTNG_BUFFER_PER_UID mode. All processes
> for our application run under one UID, and only one tracing session is
> active at a time.

Again, nothing out of the ordinary.

Next step would be a small reproducer.

Is the system under load?
Is there any cpu affinity set?
What kind of system are we talking about (x86_64? arm? 24+ cores? etc.)

Cheers

-- 
Jonathan Rajotte-Julien
EfficiOS