[omniORB] ORBscanGranularity == shutdown time?

Roloff, Craig A craig.a.roloff at boeing.com
Tue Dec 16 09:24:52 GMT 2003


On Monday 15 December, Grisby, Duncan" wrote:
> On Thursday 11 December, "Roloff, Craig A" wrote:
> 
> > Why is there a relationship between ORBscanGranularity (SG) and the
> > shutdown time?  It appears when shutdown (1) is executed that it
> > takes somewhat less than SG seconds to complete.  Here is a table
> > of SG vs shutdown times (in seconds) for a particular applicaton:
> 
> Shutdown time shouldn't be affected by the scan granularity at all.
> 
> What version of omniORB are you using?  Please try with 4.0.3. If you
> still get the delays, please run the traces again with
> -ORBtraceThreadId 1.

This behavior began once we upgraded from 3.0.4 to 4.0.3 (HP-UX 11).

The traces with -ORBtraceThreadId 1 are shown below.  Could the
delays be due to the giopServer waits?

start
=====
omniORB: (0) Preparing to shutdown ORB.
omniORB: (0) Destroying POA(RootPOA).
omniORB: (0) Deactivating all POA(RootPOA)'s objects.
omniORB: (0) Waiting for requests to complete on POA(RootPOA).
omniORB: (0) Requests on POA(RootPOA) completed.
omniORB: (0) Etherealising POA(RootPOA)'s objects.
omniORB: (0) Stopping serving incoming endpoints.
omniORB: (0) giopServer waits for completion of rendezvousers and workers

end
===
omniORB: (3) throw giopStream::CommFailure from giopStream.cc:828(0,NO,COMM_FAILURE_UnMarshalArguments)
omniORB: (3) Server connection refcount = 1
omniORB: (0) giopServer back from waiting.
omniORB: (0) Destruction of POA(RootPOA) complete.
omniORB: (0) Shutting-down all incoming endpoints.
omniORB: (0) TCP endpoint shut down.
...
omniORB: (3) Server connection refcount = 0
omniORB: (3) Server close connection from giop:tcp:136.240.18.23:61728
...
omniORB: (0) 755 object references present at ORB shutdown.
omniORB: (0) ORB shutdown is complete.
omniORB: (0) ORB not destroyed; no final clean-up.

And sandwiched between start and end are:

ORBscanGranularity = 60
=======================
omniORB: (1) AsyncInvoker: thread id = 1 has exited. Total threads = 3
omniORB: (2) Scan for idle connections (1071593545,68406000)
omniORB: (2) Scavenger reduce idle count for strand 400ccb20 to 1
omniORB: (2) Scavenger reduce idle count for strand 40427000 to 2
omniORB: (2) Scan for idle connections done (1071593545,68406000).

ORBscanGranularity = 30
=======================
omniORB: (1) AsyncInvoker: thread id = 1 has exited. Total threads = 3

ORBscanGranularity = 10
=======================
omniORB: (2) Scan for idle connections (1071592966,860368000)
omniORB: (2) Scavenger reduce idle count for strand 400ccb20 to 11
omniORB: (2) Scavenger reduce idle count for strand 40427000 to 17
omniORB: (2) Scan for idle connections done (1071592966,860368000).

ORBscanGranularity = 2
======================
<nothing>



More information about the omniORB-list mailing list