Hide Forgot
Description of problem: Upgrading RHEL 6.1 GA ppc64 Server to RHEL6.2-20110907.1 ppc64 Server via yum shows error while restarting CIM Server. Version-Release number of selected component (if applicable): RHEL6.2-20110907.1 ppc64 Server libvirt-cim-0.5.14-2.el6.ppc64 How reproducible: always Steps to Reproduce: 1. Install RHEL 6.1 GA ppc64 Server with all packages 2. Add yum repository for RHEL6.2-20110907.1 ppc64 Server 3. run yum upgrade Actual results: Updating : libvirt-cim-0.5.14-2.el6.ppc64 373/1398 Shutting down CIM server: [ OK ] [ OK ] Starting up CIM server: cimserver not started: Bind failed: Failed to bind socket on port 5989: Address already in use (error code 98). [FAILED] Updating : ... Expected results: Additional info:
I've just seen this on RHEL6.2-20110923.3 x86_64 Client too, so setting platform to All.
Hello Pavel, Would you pls kindly help to verify this bug when it's ON_QA? Since we cannot reproduce it per comment 4? Thanks, Rita
Hello Rita, I'll try to verify this bug by repeating process which detected this bug, since it happens randomly, but quite often.
Retest this bug, and i can reproduce it. Reinstall a RHEL6.1 host without the following packages: libvirt, libvirt-cim, tog-segasus ans sblim-sfcb Then yum install them manually, after all packages installed, start sblim-sfcb service first, then try to start libvirtd and tog-segasus, will get error as the bug described: # service sblim-sfcb atart # service libvirtd start Starting libvirtd daemon: [ OK ] # service tog-pegasus start tog-pegasus: Generating cimserver SSL certificates...[ OK ] Starting up CIM server: cimserver not started: Bind failed: Failed to bind socket on port 5989: Address already in use (error code 98). [FAILED] obviously, after start sblim-sfcb, port 5989 is used, and then start tog-pegasus which also need bind this port will fail.
There's a way how to reproduce it without tog-pegasus vs. sblim-sfcb conflict. It happens when cimserver is killed with SIGKILL, so it doesn't terminate cimservermain process...
Okay, my analysis of this bugs is that's a tog-pegasus problem. Basically it seems the script used on restart doesn't actually kill all the processes associated to tog-pegasus, leading to the associated port still being in use and failure to restart. I don't see how that would be specific to libvirt-cim, and unless one can provide a scenario proving libvirt-cim fault, I think it's fair to reassign the bug to tog-pegasus, Daniel
Sorry, I'm really not able to reproduce it. Script used on the restart seems to kill all the processes, when done manually (same command what is libvirt-cim executing in rpm post scriptlet), everything works fine: # while true; do /etc/init.d/tog-pegasus condrestart; done Can someone confirm that this is still happening? Pegasus was rebased between 6.1 and 6.2, maybe the issue was related to that state...
This request was not resolved in time for the current release. Red Hat invites you to ask your support representative to propose this request, if still desired, for consideration in the next release of Red Hat Enterprise Linux.
This request was erroneously removed from consideration in Red Hat Enterprise Linux 6.4, which is currently under development. This request will be evaluated for inclusion in Red Hat Enterprise Linux 6.4.
It's caused by abnormal termination. *** This bug has been marked as a duplicate of bug 957233 ***