Page 1
VMware vCenter Configuration ManagerTroubleshooting Guide VCM 5.3 This document supports the version of each product listed and supports all subsequent versions until the document is replaced by a new edition. To check for more recent editions of this document, see http://www.vmware.com/support/pubs.
Page 2
VMware products are covered by one or more patents listed at http://www.vmware.com/go/patents. VMware is a registered trademark or trademark of VMware, Inc. in the United States and/or other jurisdictions. All other marks and names mentioned herein may be trademarks of their respective companies.
Isolate the Behavior Identify External Factors Check the Debug Log Info Messages Warning Messages Error Messages Exception Messages Gather Information for VMware Customer Support Types of Problems User Interface Security / Authentication SQL Server VCM Agent UNIX Agent Report Server...
Page 4
Saving Executed Scripts and Results UNIX Agent Error Scenarios If the installation reports and error The Collector cannot ping the Agent Agent fails to return data Monitoring network traffic The Collector reports the job succeeded, but there is still no data Index VMware, Inc.
About This Book This manual, Troubleshooting Guide for VMware vCenter Configuration Manager, explains the types of problems that may occur with VMware vCenter Configuration Manager, the diagnostic material VMware Customer Supportwill need to analyze the problem, and the steps to retrieve that information.
Page 6
VCM Troubleshooting Guide VMware Professional VMware Education Services courses offer extensive hands-on labs, case study Services examples, and course materials designed to be used as on-the-job reference tools. Courses are available onsite, in the classroom, and live online. For onsite pilot...
General Troubleshooting Guidelines This document attempts to describe some of the basic steps you can take before contacting VMware Customer Support. Below is a simple flowchart that provides troubleshooting procedures, allowing you to resolve simple issues on your own. Then, if necessary, you can contact VMware Customer Support for additional support.
Administration slider. If you witness the same behavior, you then eliminate data types (one or two at a time) until you find the data type that is causing the behavior. Again, the job details on the Job History screen (as seen below) may give you some indication as to which data types failed. VMware, Inc.
Collector machine and the Agent machine. For Windows machines, Task Manager should provide you with this information (as seen in the image below). VMware, Inc.
The Collector log is the most important log file as it records the primary functionality of VCM, and when VMware Customer Support technicians speak of “the debug log” in the singular, they are typically referring to this Collector debug log. Debug logs are a common troubleshooting tool for VMware Customer Support technicians, so you may already be very familiar with the process of gathering them.
The messages contained in the logs have been specifically created to help developers understand why VCM is operating the way it is. Always review the debug logs yourself and when necessary, forward the log in its entirety on to VMware Customer Support for analysis. A sample debug log is provided below.
Gather Information for VMware Customer Support When reporting an issue to VMware Customer Support, you will save a great deal of time if you are prepared with as much information about the problem as possible. Use the “Types of Problems” section of this document to determine which diagnostic files VMware Customer Support will need in order to analyze the behavior.
UI-related problems. Screenshots typically provide the most helpful information when relating these problems to VMware Customer Support. Below is an example of a UI problem where the headers of a report are compressed into only a two or three character width.
SQL Server logs, the system and application Event Logs, and the VCM debug log are the most helpful sources of information for VMware Customer Support. Below is one example of an error you may see associated with SQL Server.
For UNIX Agent errors a debug log, UNIX Agent debug log, UNIX Agent ZRP files, and a system log are all helpful information to VMware Customer Support technicians when troubleshooting UNIX Agent issues. Below is a typical UNIX error shown in the debug log.
Report Server itself, and not with VCM. If the Report Server interface appears without error, the cause is likely in VCM. For these errors, screenshots, any support-requested SQL profiler trace files, and a debug log are the files the VMware Customer Support technicians should see. Internet Information Services (IIS) The VCM UI is displayed using Microsoft’s Internet Information Server (IIS).
Refer to VCM documentation to see the recommended hardware configurations needed to run VCM properly in the different possible environments. VMware, Inc.
Debug logs are the primary way for our developers to trace the behavior of VCM. The most important debug log is the Collector debug log, and must be extracted from the database using the EcmDebugEventViewer.exe file located in <drive>:\Program Files\VMware\VCM\Tools. The following procedure is used to extract the Collector debug log complete with INFO messages: 1.
19. Restart the VCM Collector service in the Services manager. 20. Now run the job in question again. You must wait 5 minutes after it completes before proceeding. 21. Minimize the console and navigate to <Drive>:\Program Files\VMware\VCM\Tools. In this directory, double-click ECMDebugEventViewer.exe.
4.0 key and select New | Key. Name the key Agent. 4. Select the Agent key named AreResultsSaved. Verify the value is 1. If the value does not exist, create a new DWORD value named AreResultsSaved and set it to 1. VMware, Inc.
The Event logs are divided into 3 categories: Application, Security, and System. VCM errors will almost always fall under either Application or System. Use the Event Viewer to view these logs and to export them for review by VMware Customer Support. VMware, Inc.
5. Click the Action menu and select Save Log File As. Type an appropriate name and click Save. System Information (msinfo32.exe) Occasionally,VMware Customer Support may ask for an msinfo32.exe file. Msinfo32.exe is an application that provides a fairly accurate and detailed snapshot of a system’s current activity.
Page 24
VCM Troubleshooting Guide 9. On the collector navigate to <drive>:\Program Files\VMware\VCM\SUM\Collector and look for the following files: CSISUMWorker_SumDBDebug.txt CSISUMSvc_SumDBDebug.txt CSISUMSvc_debug.txt 10. Gather all these “.txt” files together. VMware, Inc.
Troubleshooting Software Provisioning Repositories The following are problems you may encounter when working with software provisioning software repositories. Possible solutions are also provided. VMware, Inc.
1. Remove the package entries from the deleted from the repository. <path>\VMware\Tools\Repository\.hive\repository.index file. 2. Reindex all platforms and sections from which you removed the entry. To reindex, run C:\Repository>"C:\Program Files\VMware\VCM\Tools\Package Studio\hive.exe" reindex "C:\Repository\dists\Any\Release\binary-<platform>" <platform> <section>. 3. Using Package Studio to republish the package to the repository.
Possible Cause Remediation The file was manually edited. Reindex all platforms and sections from which you removed the entry. To reindex, runC:\Repository>"C:\Program Files\VMware\VCM\Tools\Package Studio\hive.exe" reindex "C:\Repository\dists\Any\Release\binary- <platform>" <platform> <section>. Too many users adding new platforms and sections Possible Cause Remediation All users have permission to add Disable the write permission to the repository.toc file for...
Incorrect path to the repository. When reindexing or publishing, use the fully qualified path to the root of the repository. For example, when publishing a package: C:\Repository>"C:\Program Files\VMware\VCM\Tools\Package Studio\hive.exe" publish "c:\Repository\internetexplorer_8.0_x86.crate" crates\i <platform> {<section>}. Troubleshooting Package Studio The following are problems you may encounter when working with the Package Studio. Possible solutions are provided.
The Generate button is grayed out Possible Cause Remediation The required fields are not yet Add valid data to the required fields, located on the populated with valid data. Manage Packages | Properties tab. The required fields are Name, Version, Architecture, Description. VMware, Inc.
Provides package name. For example, <Depends Name="internetexplorer" CrateOperator="LaterOrEqual" CrateVersion="7.0" /> is modified to read <Depends Name="InternetExplorer" CrateOperator="LaterOrEqual" CrateVersion="7.0" />. 6. Save the control.xml. 7. Save and close the .zip file. 8. Rename the .zip to .crate. VMware, Inc.
Pre-Command, Commands, Arguments, and Post-Command options. Files are added to the Project Data Directory, but they are not displayed in the list Possible Cause Remediation Display needs to be refreshed. Click the Refresh Files List button. VMware, Inc.
General Settings data grid appears. 2. Select Type of information that should be logged, and then click Edit Settings. The Edit Settings page of the General Settings wizard appears. 3. Select the following options: Exception, Error, and Info. VMware, Inc.
Page 33
<path>\CMAgent\CommonPy\AgentBridge.py elif o in ('--parameter='): param = str(a).strip("\"'") paramList = param.split('=', 1) key = paramList[0].strip("\"'") value = '' if len(paramList) > 1: value = paramList[1].strip("\"'") self.parameters[key] = value if key == 'LoginFile': from shutil import copy copy(value, 'c:\\myTemp\\') VMware, Inc.
“By Domain” or “By Machine Group” with each targeted machine. The Collector allows multiple accounts to be specified in priority order. During installation the Collector cycles through these accounts as needed to establish communication with each target machine. VMware, Inc.
Page 36
VCM Troubleshooting Guide Install The following describes in detail the install process from start to finish. Upon issuing an installation request, looking at the Jobs Running dialog box, the following jobs are listed in the order that they are executed. VMware, Inc.
Page 37
A connection is made to the module installer. b. The runtime agent lock is updated. c. The module installer returns all modules currently installed. d. The product modules are then recorded in the database. e. A failure here will cause install to fail. VMware, Inc.
Page 38
Collector. In all other cases the lock file is zeroed and the install is allowed to continue. e. A pre-installation check is then preformed to ensure that the subsequent install will succeed. If various components (ComSocketServiceListener, debug event dll, or the subsystem singleton dll) are located, they are removed. VMware, Inc.
Page 39
The module installer is pushed out and executed by the simple installer. d. The next step is to determine if any install infrastructure modules need to be upgraded. A list is made of those outdated components and they are upgraded. VMware, Inc.
Page 40
Detailed information lies outside the scope of this document. 13. Check if Request is Complete a. The Collector checks to see if the machine environment collection is complete. b. Detailed information lies outside the scope of this document. VMware, Inc.
Page 41
This step does not clear the action type in the lock. This partial release has the effect of rendering the Agent incapable of doing useful work, since this condition is a result of an invalid install. Only if step 11 is executed is the agent available for collections. VMware, Inc.
Page 42
It then in turn uninstalls each module it finds. The registration service ensures that all VCM files are removed from the file system and that the registry is purged. It then marks itself for deletion upon the next reboot of the computer. VMware, Inc.
Page 43
VCM files that it finds base on a filter stored in the database. c. Uninstall then tries to connect to the remote registry. If it can, it removes all the VMware registry entries it finds.
Wise Installation Systems 9 is the manual install program. Regardless of how it is executed, it is the program that lays down the Agent. This executable may be used interactively or silently. The following describes each usage. VMware, Inc.
CsiWin32SocketListener.exe. This module contains the http service that listens to Collector requests on a given port. The service is thin and delegates request to components that are part of the standard install. The Collector is responsible for determining the protocol to talk to the agent with. An HTTP agent can be VMware, Inc.
Page 46
Agent. Each resolution algorithm will dump the current and requested Agent modules, module dependency graph and the module updates mapping prior to algorithm execution as well as each algorithms step-by-step output as it is executes. Example debug (INFO) output captured during an agent install: VMware, Inc.
The Agent sub-directory contains code libraries that are specific to the agent. It contains a sub-directory for the code version (3.0) which in turn contains a lib directory with all libraries and a bin directory with all executables. /opt/CMAgent/Agent: VMware, Inc.
Page 48
/opt/CMAgent/CFC: drwxr-x--- 4 root cfgsoft 96 May 6 16:47 3.0 -rw-r----- 1 root cfgsoft 2677 May 6 16:47 manifest_CFC.3.0.Linux /opt/CMAgent/CFC/3.0: drwxr-x--- 2 root cfgsoft 312 May 6 16:47 bin drwxr-x--- 2 root cfgsoft 1888 May 6 16:47 lib VMware, Inc.
Page 49
152 May 8 12:28 db drwxrwx--- 4 csi_acct cfgsoft 152 May 7 11:53 tmp The db directory contains sub-directories for the data model and status manager Birdstep databases. It also contains a sub-directory for the agent's certificate store /opt/CMAgent/data/db: VMware, Inc.
Page 50
8 12:28 SM.taf -rw-rwx--- 1 root cfgsoft 268800 May 8 12:19 StateMachineStateHistoryLog.dbd -rw-rwx--- 1 root cfgsoft 41984 May 8 12:19 StateMachineStateHistoryLog_K1.dbd -rw-rwx--- 1 root cfgsoft 12800 May 8 12:19 StateMachineStatusLog.dbd -rw-rwx--- 1 root cfgsoft 2048 May 8 12:19 StateMachineStatusLog_K1.dbd VMware, Inc.
Page 51
6 16:47 lib drwxr-x--- 2 root cfgsoft 184 May 6 16:47 registration drwxr-x--- 2 root cfgsoft 376 May 6 16:47 scripts There are three files in the bin directory that are used when running inspections and remote commands. VMware, Inc.
Page 52
6 16:47 libCsiPpxLibHelper.so -r-xr-x--- 1 root cfgsoft 360720 May 6 16:47 libEcmAgentInspectorCommon.so -r-xr-x--- 1 root cfgsoft 64352 May 6 16:47 libEcmAgentInspectorScript.so -r-xr-x--- 1 root cfgsoft 274360 May 6 16:47 libEcmFileUploadJob.so -r-xr-x--- 1 root cfgsoft 226240 May 6 16:47 libEcmRemoteCommandJob.so VMware, Inc.
Page 53
The BootStrapInstall.log file contains a log of all of the actions that the installer took. The DebugEvent_cis.dbe is an error log file and can be copied to a collector machine to be viewed using the Debug Event Viewer. VMware, Inc.
Page 54
5 08:36 UserDict.py -rw-r----- 1 root cfgsoft 11066 May 6 16:47 UserDict.pyc -r--r----- 1 root cfgsoft 22409 May 5 08:36 codecs.py -rw-r----- 1 root cfgsoft 27697 May 6 16:47 codecs.pyc -r--r----- 1 root cfgsoft 6433 May 5 08:36 copy_reg.py VMware, Inc.
Page 55
Solaris agent, this directory will also contain the libiconv.so.2.1.0 library which is also world readable as it is used by the gawk executable. /opt/CMAgent/ThirdParty/1.0/bin: -r-xr-x--- 1 root cfgsoft 4376 May 6 16:47 VMwareFingerPrint -r-xr-x--x 1 root cfgsoft 308956 May 6 16:47 gawk VMware, Inc.
If the Agent was installed from the a different collector, the certificate file must be copied to the agent (e.g. via ftp in binary mode). The certificate file can be located in "C:\Program Files\VMware\VCM\Collector Data" and will be name <ECM_Enterprise_Certificate_GUID>.pem...
Agent for each collector that send requests to the Agent (for example, for a Solaris 10 x86 Agent, .PLS files for Solaris Sparc 8, 9, and 10 would also be sent to the Agent). VMware, Inc.
4096 May 20 14:12 Package The Master directory contains the inspected data and is used when performing deltas. There is 1 .mfl file per data class collected. If these files are deleted, the next collection will act as a full collect. VMware, Inc.
The types are: script_XXXXXX – a gawk or remote command script hds_XXXXXX – the output of a gawk script rcmd_XXXXXX – the output of a remote command /opt/CMAgent/data/ScriptFiles: total 8 drwxrwx--- 2 nobody cfgsoft 4096 May 20 14:08 . VMware, Inc.
9. You can use a machine that has nmap installed (usually a linux machine) to see if the port is open to the network, using the command nmap –sT –v –p 26542 <agent-machine> which will return a display like: Starting nmap V. 3.00 ( www.insecure.org/nmap/ ) Host <agent-machine> (<ip-address>) appears to be up … good. VMware, Inc.
On Red Hat Enterprise 2.1, you will see many instances of these processes. This is because this version of the OS reports all threads within a process and not just the process itself. On all other platforms, you should see only one Agent and KSrDaemon process. VMware, Inc.
Page 62
/opt/CSI/data/<collector-name>/<request-id>. This file will eventually be returned to the collector. It can be accessed on the agent as described in As with the Windows Agent, many of the errors in the dbe files require intimate knowledge of the Agent code in order to decipher them. VMware, Inc.
In very rare cases, it may be useful to monitor the tcp traffic for the agent. The easiest way to do this is if you have X windows access to the client and the ethereal package is installed. VMware, Inc.
Page 64
As messages flow in and out of the Agent’s port (26542), they will appear in the ethereal display. By looking at the colored messages in the display, it is possible to determine what messages have occurred. For a single inspection, expect to see the following: VMware, Inc.
Page 65
/usr/sbin/tcpdump –s –l 256 –x –X port 26542 | tee tcpdump.log The tcpdump.log file contains similar information as described for ethereal except that there is no color highlighting, making it necessary to search for the string HTTP to determine where each of the different message types begins. VMware, Inc.
Also, the mount options for the file system should be checked. A common security practice is to mount /usr, /opt, and /usr/local with 'notsetuid/nosuid' options to prevent setuid binaries from running. This will prevent RunHigh/RunLow/RunRemote from functioning. VMware, Inc.
Need help?
Do you have a question about the VCM 5.3 and is the answer not in the manual?
Questions and answers