Oracle eMail Server Administrator's Guide Release 5.2 A86653-01 |
|
eMail Server provides several tools and reports to assist you in monitoring and troubleshooting your messaging system. Using the Administration Tool or OOMGR, you can monitor the operation of your eMail Server system. These tools offer a wide variety of tests that you can run automatically to monitor message flow and database space usage and report any system problems.
This chapter contains the following information on using those tools and generating reports:
The eMail Server server process logs provide a continuously running account of system operations and events. They contain entries for all normal operations and for all errors that occur. Log files are useful for monitoring system performance, but unlike monitor reports, the information is not sent to a user account so you must check the log files periodically to identify problems.
All log files are located in $ORACLE_HOME/office/log /
<node_ sid>.
The log filename format is <hostname_server>_<process_name><instance_number>.log, where hostname_server is the name of the host computer for the database, process_name is the process type, and instance_number is the number of the process instance you're checking.
Following are examples of the process log file names for instance 1 of each process. The database has the SID acme and is located on the acmehost computer.
Most processes have parameters where you can specify the level of logging you want to see in the log files. eMail Server provides standards for log levels, but keep in mind that these standards are applied to the processes in different ways. For example, some processes may require only the first two log levels, or two processes may display different types of information displayed for log level 3 (medium information).
Log Level | Description |
---|---|
|
No logging |
|
Error messages only |
|
Minimum information and warnings |
|
Medium information |
|
Maximum information |
|
Debugging information |
See Also:
Chapter 11, "Process Parameter Reference", for more information about setting the log level parameter for a specific process |
eMail Server provides standards for reporting messages in log files so that you can easily find and interpret the information you need to monitor your system effectively.
The standards are different for different types of messages, but all messages contain a time stamp and type code. The type code indicates to which log level the message belongs. You can use this information to determine whether the log level for a parameter is displaying the type of information you want to see. The type codes are as follows:
Type Code | Log Level | Description |
---|---|---|
|
|
Error messages only |
|
|
Minimum information and warnings |
|
|
Medium or Maximum information |
|
|
Debugging information |
For example, if the log level for a postman process is set to 4, the log file will display messages containing the ERR
, INF
, and DIAG
codes.
Log messages reporting normal operations contain the date, time, type code, and description of the log entry.
For example:
11/16 18:13:23 INF: Guardian process started.
Error messages that appear in the log files contain the date, time, type code, component ID, error number, and description of the log entry.
For example:
11/16 18:13:23 ERR ORA-942: Table or view does not exist
To display a cause and action for this error message, you can enter the following command at the command line (make sure ORACLE_HOME is set correctly first):
$ oerr ora 942
The error number (in this example, ORA-942) corresponds to the errors listed in Chapter 19, "Error Codes and Messages". The component ID indicates which server process experienced the error.
Log messages that reflect user actions for the protocol server processes such as the POP3SRV process and the IMAP4SRV process contain the date, time, type code, user ID, thread ID, description of the log entry.
For example:
11/16 18:13:23 INF jdoe.10: Login succeeded
eMail Server provides a monitoring pack that tracks the capacity of the system through Oracle Enterprise Manager (OEM). The monitoring pack displays the length of message queues and number of concurrent IMAP server connections in a series of charts.
To use the eMail Server capacity monitoring pack, you must first configure your system.
lsnrctl dbsnmp_stop
IOFCMGR> startup all:
lsnrctl dbsnmp_start
eMail Server provides a default user account called ORAPOST
that receives error messages from the monitor reports and statistic tasks and notifications regarding messages that could not be sent. This is a standard user account, so you can check the messages using the same client software that your users use to check their messages. You can also change the recipient account by modifying certain process parameters.
Monitor reports identify problem areas and suggest ways to avoid possible problems, such as low disk space, before they occur. You can choose from a wide variety of monitoring reports to help you keep your system running smoothly.
You select tests from the list provided, and specify how often to run them. Once the tests are run, the monitor sends a report. Monitor reports are sent as e-mail messages to an account that you specify (the default is ORAPOST
). You can save these messages, write them to a file, print them, or do anything with them that you can do with an e-mail message.
See Also:
|
Messages from the monitor and statistics processes are directed to a user account in your directory called ORAPOST
. If there are no problems, the monitor sends a "No Problem" report to this account. A No Problem report is a blank message with a subject line that you specify in the goodSubj
parameter for the monitor process. If a test discovers a problem or potential problem, the monitor sends a "Problem" report to this account. Problem reports contain information about any problems found, as well as suggestions for fixing the problems. Problem reports consist of subreports that correspond to particular tests.
The ORAPOST
account also receives messages from the postman process.
You can change the user account to which this information is sent by modifying the postmaster
parameter for the postman process, or the probRecips
and noProbRecips
parameters for the monitor and statistics processes.
You can view the ORAPOST
account by logging on to the system with a messaging client as you would for any other user. You can change the password for the ORAPOST
account by changing the password
attribute for the ORAPOST
directory entry. Refer to "Setting a User's Password" for instructions.
You must run the monitor and statistics processes to collect data and run the tests used to create reports. The monitor process checks message flow and database space usage, and the statistics process collects information about delivery time and database space usage that is used by the monitor tests to create reports.
You can start the monitor and statistics processes and run them at configurable intervals, or you can use the following instructions to see immediate results in the Administration Tool.
See Also:
"Starting a Registered Process" for more information about the configurations of the monitor and statistics processes |
This task can only be performed through the Administration Tool GUI.
The results appear in the Diagnose dialog box. The Test Fail? column contains an X for any tests that failed.
After selecting the tests you want to run, you must activate the tests. Once activated, the tests run automatically whenever the monitor process is running.
In the Administration Tool, you must run the statistics tasks and monitor tests once before you can activate a monitor test. Refer to "Running the Monitor Tests and Statistics Tasks" for instructions.
This task can be performed through either the Administration Tool GUI, or the OOMGR command-line interface.
IOFCMGR>modify monitor name=<test_name
> to active=Y;
You can specify which tasks you want the statistics process to perform and how often these tasks are performed.
Note: The statistics process only gathers data for the tests that are active. Refer to "Activating the Monitor Tests" for instructions on activating tests. |
This task can only be performed through the OOMGR command-line interface.
IOFCMGR>modify statistics name=<task_name
> 2>to active=Y frequency=<task_performance_interval_in_minutes
>;
IOFCMGR>modify statistics name=<task_name
> to active=N;
The statistics process can perform the following tasks:
You can deactivate a test if you no longer want it to run automatically.
This task can be performed through either the Administration Tool GUI, or the OOMGR command-line interface.
om_mon_test
table. The Active
column in this table contains a Y
if the test is active.
IOFCMGR>modify monitor name=<test_name
> to active=N;
In addition to the information available through the database space tests, you can also check all tablespaces on a given node. This procedure displays the bytes used, bytes free, and the maximum free extents for each tablespace.
This task can only be performed through the OOMGR command-line interface.
In addition to the information available through the database space tests, you can also check the space used by an individual user.
This task can only be performed through the OOMGR command-line interface.
Protocol server tests probe the listener to retrieve run-time information about the protocol server processes. You can use this information to determine whether eMail Server is configured for optimal performance. For example, if you notice that some of the database connections are not being used regularly, you can lower the minimum number of connections for the protocol server and reduce the memory usage.
See Also:
"Understanding the Probe Results" for information about how to interpret the listener probe results |
Before running the listener probe, determine what kind of information you want to see. The following information is available:
This task can be performed through either the Administration Tool GUI, or through a telnet session.
You can click this button whenever you want to refresh the information in the right pane.
$ telnet <server
> <diagport
>
The default diagport
is 5010
.
You should see something like the following:
Trying 111.11.1.111... Connected to <server
>. Escape character is '^]'. usage: probe <class_ID
> <instance_ID
> [action=dbinfo | clientinfo | all] probe 0 0 action=shminfo
See Also:
"Parameters for Monitoring Protocol Server Processes" for more information about the parameters available with this command |
Use the following values when running the Probe through a telnet session:
Value | Description |
---|---|
|
Valid values: |
|
Instance number of the protocol server process you want to check |
The following example shows typical results of the probe
command for displaying database connection information (probe <
class_ID
> <
instance_ID
> action=dbinfo
).
Tue May 19 11:53:34 1998 Database: msgdata Con Audsid St Access# T# F# UserName.ThreadId XCur FCur --- -------- -- ------- -- -- ----------------- ---- ---- 0 18275860 0 3875 0 0 llane5.43 1 18275862 5 1154 1 0 ckent1.12 112 2 18275863 1 398 0 1 jolsen2.143 27 3 18275864 0 4 0 0 bwayne1.12 4 18275866 0 0 0 0 5 18275867 0 0 0 0 6 18275868 0 0 0 0 7 18275869 0 0 0 0 8 18275870 0 0 0 0
These results show that there are eight connections to the msgdata
database. It also shows runtime usage information about these connections.
The following parameters appear in the listener probe output for database connection information (probe <
class_ID
> <
instance_ID
> action=dbinfo
).
The following parameters appear in the listener probe output for client connection information (probe <
class_ID
> <
instance_ID
> action=clientinfo
).
The following parameters appear in the listener probe output for shared memory information (probe 0 0 action=shminfo
).
If a client is disconnected from the e-mail server unexpectedly, the user's INBOX may be locked so the user cannot access it after logging in again.
This task can only be performed through SQL*Plus.
OO
.
SYS
.
The script tells you what SQL statement to use to unlock the INBOX. In the following example, 11 represents the SID, and 2474 is the serial number.
ALTER SYSTEM KILL SESSION '11, 2474';
ALTER SYSTEM KILL SESSION '11, 2474';
). This removes the lock so the user can access the INBOX.
|
Copyright © 2001 Oracle Corporation. All Rights Reserved. |
|