Bug 470167

Summary:	ManagementAgent::Singleton destructor aborts
Product:	Red Hat Enterprise MRG	Reporter:	Matthew Farrellee <matt>
Component:	qpid-qmf	Assignee:	Ted Ross <tross>
Status:	CLOSED ERRATA	QA Contact:	Jeff Needle <jneedle>
Severity:	urgent	Docs Contact:
Priority:	urgent
Version:	1.0	CC:	freznice, tross
Target Milestone:	1.1
Target Release:	---
Hardware:	All
OS:	Linux
Whiteboard:
Fixed In Version:		Doc Type:	Bug Fix
Doc Text:		Story Points:	---
Clone Of:		Environment:
Last Closed:	2009-02-04 15:35:51 UTC	Type:	---
Regression:	---	Mount Type:	---
Documentation:	---	CRM:
Verified Versions:		Category:	---
oVirt Team:	---	RHEL 7.3 requirements from Atomic Host:
Cloudforms Team:	---	Target Upstream Version:
Embargoed:
Bug Depends On:
Bug Blocks:	469388

Description Matthew Farrellee 2008-11-06 04:09:56 UTC

Using qpid trunk from svn, r711740.

This issue was originally reported: https://bugzilla.redhat.com/show_bug.cgi?id=469388

The code failing for BZ469388 did not shutdown the ManagementAgent::Singleton before exiting. The new code does shut it down, by delete'ing it, and still the daemon aborts.


Code in condor_master...

ManagementAgent::Singleton *singleton;

void
initialize()
{
...
   singleton = new ManagementAgent::Singleton();
...
}

void
shutdown()
{
   dprintf(D_ALWAYS, "MgmtMasterPlugin shutting down...\n");
   dprintf(D_ALWAYS, "...%p\n", singleton);
   if (singleton) {
      dprintf(D_ALWAYS, "...deleting ManagementAgent\n");
      delete singleton;
      singleton = NULL;
      dprintf(D_ALWAYS, "...deleted ManagementAgent\n");
   }
}

Log showing the master abort instead of shutting down cleanly...

11/5 21:57:03 All daemons are gone.  Exiting.
11/5 21:57:03 MgmtMasterPlugin shutting down...
11/5 21:57:03 ...0x85bc7a8
11/5 21:57:03 ...deleting ManagementAgent
Invalid argument
condor_master: /usr/local/include/qpid/sys/posix/Mutex.h:116: void qpid::sys::Mutex::lock(): Assertion `0' failed.
Stack dump for process 18099 at timestamp 1225943823 (12 frames)
./release_dir/sbin/condor_master(dprintf_dump_stack+0xa6)[0x80f095c]
./release_dir/sbin/condor_master[0x80f0b56]
[0x110400]
/lib/libc.so.6(abort+0x101)[0x52bf91]
/lib/libc.so.6(__assert_fail+0xee)[0x52393e]
/home/matt/Documents/Condor/Installation/libexec/MgmtMasterPlugin-plugin.so(_ZN4qpid3sys5Mutex4lockEv+0x60)[0x25b62c]
/usr/local/lib/libqpidclient.so.0(_ZN4qpid6client10Dispatcher3runEv+0x92c)[0x2d62ac]
/usr/local/lib/libqpidclient.so.0(_ZN4qpid6client19SubscriptionManager3runEv+0x35)[0x308385]
/usr/local/lib/libqmfagent.so.0(_ZN4qpid10management19ManagementAgentImpl16ConnectionThread3runEv+0x7b5)[0x347975]
/usr/local/lib/libqpidcommon.so.0[0xf58bc1]
/lib/libpthread.so.0[0x69350b]
/lib/libc.so.6(clone+0x5e)[0x5d4b2e]
zsh: abort      ./release_dir/sbin/condor_master -t -f

Comment 1 Ted Ross 2008-11-06 18:32:59 UTC

To verify this fix, use the qmf-agent example.  You are looking for a clean shutdown when you ^C the program (no SEGVs or unhandled exceptions).

Run the example with no broker.  Verify shutdown.

Start a broker and run the example.  Verify shutdown.

Run the example on a broker, stop the broker, then verify shutdown.

The shutdown of qmf-agent should be clean in all cases.

Comment 3 Frantisek Reznicek 2008-11-20 08:47:59 UTC

RHTS test qpid_test_qmf_agent_bz470167 validates that this issue has been fixed.
->VERIFIED

Comment 5 errata-xmlrpc 2009-02-04 15:35:51 UTC

An advisory has been issued which should help the problem
described in this bug report. This report is therefore being
closed with a resolution of ERRATA. For more information
on therefore solution and/or where to find the updated files,
please follow the link below. You may reopen this bug report
if the solution does not work for you.

http://rhn.redhat.com/errata/RHEA-2009-0035.html