Description of problem: Uninventory of platform complains about missing resource key and fails to delete spool for deleted agent. I noted this while verifying Bug 655187 Version-Release number of selected component (if applicable): jon-server-2.4.1-SNAPSHOT build# ae99b5b How reproducible: Always. Steps to Reproduce: 1. Inventory a platform. 2. Uninventory the platform. 3. Monitory server log. Actual results: 2010-11-25 14:56:07,714 WARN [org.hibernate.hql.ast.QueryTranslatorImpl] firstResult/maxResults specified with collection fetch; applying in memory! 2010-11-25 14:56:17,658 INFO [org.rhq.enterprise.server.resource.ResourceManagerBean] User [org.rhq.core.domain.auth.Subject[id=2,name=rhqadmin]] is marking resource [Resource[id=10001, type=Linux, key=RHEL 5.5, name=RHEL 5.5, parent=<null>, version=Linux 2.6.18-194.17.1.el5]] for asynchronous uninventory 2010-11-25 14:56:18,028 WARN [org.rhq.enterprise.server.core.comm.ServerCommunicationsService] {Failed to truncate/delete spool for deleted agent [Agent[id=10001,name=RHEL 5.5,address=10.65.201.129,port=16163,remote-endpoint=socket://10.65.201.129:16163/?rhq.communications.connector.rhqtype=agent&numAcceptThreads=1&maxPoolSize=303&clientMaxPoolSize=304&socketTimeout=60000&enableTcpNoDelay=true&backlog=200,last-availability-report=1290677005767]] please manually remove the file: null}!!! missing resource message key=[Failed to truncate/delete spool for deleted agent [Agent[id=10001,name=RHEL 5.5,address=10.65.201.129,port=16163,remote-endpoint=socket://10.65.201.129:16163/?rhq.communications.connector.rhqtype=agent&numAcceptThreads=1&maxPoolSize=303&clientMaxPoolSize=304&socketTimeout=60000&enableTcpNoDelay=true&backlog=200,last-availability-report=1290677005767]] please manually remove the file: null] args=[java.lang.NullPointerException] 2010-11-25 14:56:18,028 INFO [org.rhq.enterprise.server.core.AgentManagerBean] Removed agent: Agent[id=10001,name=RHEL 5.5,address=10.65.201.129,port=16163,remote-endpoint=socket://10.65.201.129:16163/?rhq.communications.connector.rhqtype=agent&numAcceptThreads=1&maxPoolSize=303&clientMaxPoolSize=304&socketTimeout=60000&enableTcpNoDelay=true&backlog=200,last-availability-report=1290677005767] 2010-11-25 14:56:18,154 ERROR [org.rhq.enterprise.server.discovery.DiscoveryServerServiceImpl] Received invalid inventory report from agent [Agent[id=0,name=RHEL 5.5,address=null,port=0,remote-endpoint=null,last-availability-report=null]]: Unknown Agent named [RHEL 5.5] sent an inventory report - that report will be ignored Expected results: There should be no stray files left when uninventoried. Additional info:
I've opened separate bug (and did not update Bug 655187) as I get the above error even without bundle being uploaded.
The exception about failure to delete/truncate spool was fixed by putting a check in place that checked the existence of the spool file. The error messages about invalid inventory (or availability) reports still appear, but their content was updated to state the fact that if such error happens right after platform removal, it can be ignored. commit a6d145b4c1f2f200a9bbcfc13933a01e9815f01b Author: Lukas Krejci <lkrejci> Date: Mon Nov 29 18:35:13 2010 +0100 BZ 657232 * Don't throw NPE while destroying an agent on the server-side if there is no command spool file name defined. * Improve the error messages during intermittent confusion of the server and agent after platform removal.
Verified this against jon-server-2.4.1-SNAPSHOT build# cef77c5. I don't see the error message now. Below is the log snippet. 2010-12-01 12:13:54,529 INFO [org.rhq.enterprise.server.resource.ResourceManagerBean] User [org.rhq.core.domain.auth.Subject[id=2,name=rhqadmin]] is marking resource [Resource[id=10001, type=Linux, key=dhcp6-150.pnq.redhat.com, name=dhcp6-150.pnq.redhat.com, parent=<null>, version=Linux 2.6.18-194.17.1.el5]] for asynchronous uninventory 2010-12-01 12:13:55,002 INFO [org.rhq.enterprise.server.core.AgentManagerBean] Removed agent: Agent[id=10001,name=dhcp6-150.pnq.redhat.com,address=10.65.201.129,port=16163,remote-endpoint=socket://10.65.201.129:16163/?rhq.communications.connector.rhqtype=agent&numAcceptThreads=1&maxPoolSize=303&clientMaxPoolSize=304&socketTimeout=60000&enableTcpNoDelay=true&backlog=200,last-availability-report=1291185687763] 2010-12-01 12:13:55,143 ERROR [org.rhq.enterprise.server.discovery.DiscoveryServerServiceImpl] Received invalid inventory report from agent [Agent[id=0,name=dhcp6-150.pnq.redhat.com,address=null,port=0,remote-endpoint=null,last-availability-report=null]]: Unknown Agent named [dhcp6-150.pnq.redhat.com] sent an inventory report - that report will be ignored. This error is harmless and should stop appearing after a short while if the platform of the agent [dhcp6-150.pnq.redhat.com] was recently removed from the inventory. In any other case this is a bug. 2010-12-01 12:13:56,193 INFO [org.rhq.enterprise.server.core.CoreServerServiceImpl] Got agent registration request for new agent: dhcp6-150.pnq.redhat.com[10.65.201.129:16163][3.0.1-SNAPSHOT(cef77c5)] Marking this bug as verified.
Bookkeeping - closing bug - fixed in recent release.