[Bug 1325363] New: multithreaded clients with concurrent pmNewContext suffer deadlocks or low performance

[bugzilla@xxxxxxxxxx]

https://bugzilla.redhat.com/show_bug.cgi?id=1325363

            Bug ID: 1325363
           Summary: multithreaded clients with concurrent pmNewContext
                    suffer deadlocks or low performance
           Product: Fedora
           Version: rawhide
         Component: pcp
          Assignee: fche@xxxxxxxxxx
          Reporter: fche@xxxxxxxxxx
        QA Contact: extras-qa@xxxxxxxxxxxxxxxxx
                CC: brolley@xxxxxxxxxx, fche@xxxxxxxxxx, lberk@xxxxxxxxxx,
                    mgoodwin@xxxxxxxxxx, nathans@xxxxxxxxxx,
                    pcp@xxxxxxxxxxx, scox@xxxxxxxxxx



Created attachment 1145152
  --> https://bugzilla.redhat.com/attachment.cgi?id=1145152&action=edit
test program for forthcoming fix

While parallelizing pmmgr's scanning logic, ran into a roadblock within libpcp,
namely pmNewContext holding libpcp throughout its operation.  This means that
opens to remote PM_CONTEXT_HOST's that are unreachable or slow serialize the
client program, even if itself multithreaded.

A test case that demonstrates much better performance with a multithreaded
pmNewContext (I have a prototype) shows outright deadlocks with current libpcp,
not even just sloth.  Attaching test case.  Invoke with

  % ./multithread10 localhost localhost

-- 
You are receiving this mail because:
You are on the CC list for the bug.
Unsubscribe from this bug 
https://bugzilla.redhat.com/token.cgi?t=3XHwWswvDd&a=cc_unsubscribe