Version-Release number of selected component (if applicable): sesame-0.4.3030-2.el5 How reproducible: 100% Steps to Reproduce: 1. service sesame start 2. service sesame start 3. service sesame status Actual results: 2 sesame daemons will be running Expected results: 1 sesame daemon running, from step 1, and just a return of code 0 for step 2. Additional info: This is a problem because it results in a single machine being over represented in the qmf data stream, and confuses qmf ids with collisions.
I think the sesame daemon itself should enforce singletonness. Do you agree, Ted?
Quick crosscheck on beta1 [root@mrg-qe-02 qpid_ptest_authentication_krb5]# rpm -q sesame sesame-0.7.3918-1.el5 [root@mrg-qe-02 qpid_ptest_authentication_krb5]# service sesame status sesame is stopped [root@mrg-qe-02 qpid_ptest_authentication_krb5]# service sesame start Starting Sesame daemon: [ OK ] [root@mrg-qe-02 qpid_ptest_authentication_krb5]# service sesame status sesame (pid 4953) is running... [root@mrg-qe-02 qpid_ptest_authentication_krb5]# service sesame start Starting Sesame daemon: [ OK ] [root@mrg-qe-02 qpid_ptest_authentication_krb5]# service sesame status sesame (pid 5007 4953) is running... The issue is still present on sesame-0.7.3918-1.el5
It is still possible to run unlimited number of sesames from one host. sesame-0.7.3918-6.el5
Here is a simple patch to enforce the singleton nature from the init script. It should be sufficient to satisfy requirements with a small impact on sesame. As an side, it is valuable to be able to run many sesame daemons on a single machine. Index: cpp/etc/sysvinit-sesame =================================================================== --- cpp/etc/sysvinit-sesame (revision 4429) +++ cpp/etc/sysvinit-sesame (working copy) @@ -22,6 +22,10 @@ RETVAL=0 start() { + # Run only a single daemon + status $processname > /dev/null && exit 0 + rm -f /var/lock/subsys/$servicename + echo -n $"Starting Sesame daemon: " daemon --user sesame --check $servicename $processname $SESAME_OPTIONS\& RETVAL=$?
The issue is not yet included in last packaged sesame version (sesame-0.7.4297-4.el5). See details below: [root@mrg-qe-02 ~]# service sesame status sesame (pid 4046) is running... [root@mrg-qe-02 ~]# service sesame start Starting Sesame daemon: [ OK ] [root@mrg-qe-02 ~]# service sesame status sesame (pid 26925 4046) is running... [root@mrg-qe-02 ~]# rpm -q sesame sesame-0.7.4297-4.el5.x86_64 [root@mrg-qe-02 ~]#
Waiting for the package with fix. (s/The issue is not yet included/The issue patch is not yet included/ in comment 7)
Technical note added. If any revisions are required, please edit the "Technical Notes" field accordingly. All revisions will be proofread by the Engineering Content Services team. New Contents: C: cumin init script would not detect service already running on start C: Multiple cumin instances could be started at once F: The init script now uses a subsys lock to avoid multiple runs R: service cumin start is now idempotent
Built sesame-0.7.4297-5.el5 to include the above patch. See https://brewweb.devel.redhat.com/buildinfo?buildID=154503
sesame-0.7.4297-5.el5 fixes the problem reliably. Waiting for el4 package to be able to close this defect.
The issue has been fixed, tested on RHEL 4.8 / 5.6 i386 / x86_64 on package sesame-0.7.4297-5.el4/5 -> VERIFIED
This is not a Cumin bug, I am setting it into management component.
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/RHBA-2011-0217.html