Red Hat Bugzilla – Bug 711852
second ASEHA agent wont start - Unique attribute collision
Last modified: 2013-08-14 19:06:33 EDT
Description of problem: When using (sybase database) ASEHAagent resource in cluster.conf there is a unique attribute collision on attr=sybase_ase Version-Release number of selected component (if applicable): rgmanager-3.0.12-10.el6.x86_64 How reproducible: Steps to Reproduce: 1. In cluster.conf setup two ASEHAagent resources with the same value for the sybase_ase attribute 2. restart rgmanager 3. Jun 7 13:51:49 hostname rgmanager[3996]: Reconfiguring Jun 7 13:51:49 hostname rgmanager[3996]: Loading Service Data Jun 7 13:51:50 hostname rgmanager[3996]: Unique attribute collision. type=ASEHAagent attr=sybase_home value=/opt/sybase/ASE15 Jun 7 13:51:50 hostname rgmanager[3996]: Error storing ASEHAagent resource Jun 7 13:51:50 hostname rgmanager[3996]: Unique attribute collision. type=ASEHAagent attr=sybase_ase value=ASE-15_0 Jun 7 13:51:50 hostname rgmanager[3996]: Error storing ASEHAagent resource Jun 7 13:51:50 hostname rgmanager[3996]: Unique attribute collision. type=ASEHAagent attr=sybase_ase value=ASE-15_0 Jun 7 13:51:50 hostname rgmanager[3996]: Error storing ASEHAagent resource Actual results: Jun 7 13:51:49 hostname rgmanager[3996]: Reconfiguring Jun 7 13:51:49 hostname rgmanager[3996]: Loading Service Data Jun 7 13:51:50 hostname rgmanager[3996]: Unique attribute collision. type=ASEHAagent attr=sybase_home value=/opt/sybase/ASE15 Jun 7 13:51:50 hostname rgmanager[3996]: Error storing ASEHAagent resource Jun 7 13:51:50 hostname rgmanager[3996]: Unique attribute collision. type=ASEHAagent attr=sybase_ase value=ASE-15_0 Jun 7 13:51:50 hostname rgmanager[3996]: Error storing ASEHAagent resource Jun 7 13:51:50 hostname rgmanager[3996]: Unique attribute collision. type=ASEHAagent attr=sybase_ase value=ASE-15_0 Jun 7 13:51:50 hostname rgmanager[3996]: Error storing ASEHAagent resource Expected results: no unique attribute collision Additional info:
FYI RHEL 5 equivalent which hasn't been fixed either https://bugzilla.redhat.com/show_bug.cgi?id=690265
Before fix: [root@ask-03 ~]# rg_test test /etc/cluster/cluster.conf | grep Error Running in test mode. Loaded 24 resource rules Error: Unique attribute collision. type=ASEHAagent attr=sybase_home value=/opt/sybase Error storing ASEHAagent resource Error: Reference to nonexistent resource syb_ase2 (type ASEHAagent) After fix (with scratch rpm): [root@ask-03 ~]# rg_test test /etc/cluster/cluster.conf | grep ErrorRunning in test mode. Loaded 24 resource rules [root@ask-03 ~]#
Created attachment 509850 [details] Cluster.conf used to replicate bug and test fix
Created attachment 510751 [details] Patch to fix bz
http://post-office.corp.redhat.com/archives/cluster-list/2011-June/msg00132.html Chris packages have been tested and results show that the change works as expected.
I believe that the "server_name" parameter should remain unique.
Works as expected with reproducer config. But according to comment 5 and later also comment 10 the server_name should still be unique and it's not (see the patch). Validation with duplicate server_name still passes the tests, so I'm switching this back to assigned. Please make server_name unique.
https://www.redhat.com/archives/cluster-devel/2011-October/msg00062.html
Given the simplest possible cluster.conf that would exhibit this behavior: [root@snap ~]# cat sybasetest.conf <cluster> <rm> <resources> <ASEHAagent name="a" sybase_home="b" sybase_ase="c" sybase_ocs="not_unique" server_name="not_unique_test" login_file="a" interfaces_file="a" sybase_user="not_unique" shutdown_timeout="120" start_timeout="120" deep_probe_timeout="120" /> <ASEHAagent name="b" sybase_home="b" sybase_ase="c" sybase_ocs="not_unique" server_name="not_unique_test" login_file="a" interfaces_file="a" sybase_user="not_unique" shutdown_timeout="120" start_timeout="120" deep_probe_timeout="120" /> </resources> </rm> </cluster> Prior to patch (3.9.2-6.el6): === Resources List === Resource type: ASEHAagent Agent: ASEHAagent.sh Attributes: name = a [ primary unique ] sybase_home = b [ required ] sybase_ase = c [ required ] sybase_ocs = not_unique [ required ] server_name = not_unique_test [ required ] login_file = a [ required ] interfaces_file = a [ required ] sybase_user = not_unique [ required ] shutdown_timeout = 120 [ required ] start_timeout = 120 [ required ] deep_probe_timeout = 120 [ required ] Resource type: ASEHAagent Agent: ASEHAagent.sh Attributes: name = b [ primary unique ] sybase_home = b [ required ] sybase_ase = c [ required ] sybase_ocs = not_unique [ required ] server_name = not_unique_test [ required ] login_file = a [ required ] interfaces_file = a [ required ] sybase_user = not_unique [ required ] shutdown_timeout = 120 [ required ] start_timeout = 120 [ required ] deep_probe_timeout = 120 [ required ] === Event Triggers === Event Priority Level 100: Name: Default (Any event) File: /usr/share/cluster/default_event_script.sl After patch (corrected behavior): Error: Unique attribute collision. type=ASEHAagent attr=server_name value=not_unique_test Error storing ASEHAagent resource Loaded 26 resource rules === Resources List === Resource type: ASEHAagent Agent: ASEHAagent.sh Attributes: name = a [ primary unique ] sybase_home = b [ required ] sybase_ase = c [ required ] sybase_ocs = not_unique [ required ] server_name = not_unique_test [ unique required ] login_file = a [ required ] interfaces_file = a [ required ] sybase_user = not_unique [ required ] shutdown_timeout = 120 [ required ] start_timeout = 120 [ required ] deep_probe_timeout = 120 [ required ] === Event Triggers === Event Priority Level 100: Name: Default (Any event) File: /usr/share/cluster/default_event_script.sl
Note both the error storing the second resource with the non-unique value and also the change in flags from [ required ] to [ unique required ]
resource-agents-3.9.2-7.el6.x86_64 works as expected now, server_name unique.
Since the problem described in this bug report should be resolved in a recent advisory, it has been closed with a resolution of ERRATA. For information on the advisory, and where to find the updated files, follow the link below. If the solution does not work for you, open a new bug report. http://rhn.redhat.com/errata/RHSA-2011-1580.html