Tom Tucker

@morrone, to your point, I'm not disinclined to remove that function and roll it into some other API, for example, **ldmsd_set_new**, I'm just explaining how it got there.

Is there general consensus that we should convert our logging from 'level' to a bit mask with a bit for each type of error?

@baallan, I believe the missing lock here is intentional. The set->lock is a hot-lock to the extent that it will hold off set updates and other important latency sensitive activities...

Hi Narate, I think this may have to do with the fact that there is already an open connection on one of the threads that would become invalid after the...

Pfft, Eric is way ahead of me... On Tue, Aug 24, 2021 at 2:46 PM Eric Roman ***@***.***> wrote: > This isn't quite a fix, but I did try moving...

But my speculation is that it's not the threads, it's the connection that owns resources that are not available in the new namespace... On Tue, Aug 24, 2021 at 2:53...

Where does the above program fail? The call to setns()? On Tue, Aug 24, 2021 at 3:21 PM narategithub ***@***.***> wrote: > Sure. I'll try it. But closing the transport...

The *setns*() system call allows the calling thread to move into different namespaces. The *fd* argument is one of the following: • a file descriptor referring to one of the...

Do we know which event is preceded by the call to setns()? On Tue, Aug 24, 2021 at 4:04 PM narategithub ***@***.***> wrote: > The file is a soft link...

Sorry, I mean precedes. Which event occurs before the call to setns() by slurm? The reason I'm asking is that we can defer these events and then deliver them _after_...