Shared memory not getting cleaned up

[aingerson]

Using Open MPI v4.0.0 compiled from source on RHEL7.4

---

Details of the problem

I've already let @rhc54 know about this, but figured I'd add this here for record keeping/tracking.

On our CI testing of libfabric + Open MPI, I've noticed a lot of "vader_segments" in /dev/shm that are accumulating from all our testing. I think this is related to issue #5798 and was fixed with commit c076be5 but I don't think this patch was pulled into 4.0.x. If this was fixed in a different version, which version and sorry about the spam!

[jsquyres]

@gpaulsen @hppritcha FYI. This may end up being a dup of #6565, but #6565 has much more detail.

[jsquyres]

@aingerson Thanks for the research to find c076be5. I've put this information on #6565; even though that one was filed later, it has additional detail. So I'm going to close this one as a dup -- let's keep following on #6565.

[leofang]

Also dup: #6322