[Enterprise Extensions only]

Reading a formatted activity log

The WebSphere Application Server enterprise services' activity logs and trace logs must be formatted before you can read their contents. You can use the showlog utility to format an activity log or trace log, and pipe the formatted output to a file as described in Formatting an activity or trace log.

To read a formatted activity log, you can use you can use a text editor

It is easier to locate the cause of a problem in smaller activity logs. Therefore, consider reducing the size of the activity log before attempting to read it. For more information about, see Hints and tips: activity log.

When reading a formatted activity log, you need to identify the group of entries that are related to the problem or error that you want to resolve. A group of entries forms a bracket, as follows:

The start of the bracket
Initial failure, which is a single entry in the log
Results of the initial failure
A number of entries in the log
The end of the bracket
Last result of the failure, which is a single entry in the log

In general, when you are reading the activity log, you start with its last entry and then work backwards, reading the previous entry, then the one before that, and so on.

To find the bracket of entries for a problem that you are diagnosing, complete the following steps:

  1. Identify the end of the bracket.
    The first step in reading the activity log is to identify the last entry that reported the problem that you want to diagnose (that is, the end of bracket entry). This is essential for identifying the cause of the problem. You want to start with the latest entry in the activity log and search backwards for the entry that reports the problem.
    Note: Sometimes, the last entries of the log do not relate to the problem that you are interested in. For example, there may be entries made by the ORB or by requests not associated with the problem

    If you do not know the UOW for the problem that you want diagnose, you can examine the entries by alternative groupings, such as TimeStamp. You can look for the end of bracket by searching in the log starting with its last entry with the TimeStamp.

    When the entry related to the failure has been identified, you have found the end of the bracket. Remember the unit of work (UOW) and record IDs (Rec_nnnn) for the end of bracket entry. In the next step, you look at entries within that unit of work.

    Sometimes you may not have any UOW identification for the end of bracket entry. In such situations, you must look at entries that do not have a UOW identification.

  2. Find relevant entries.
    Examine each entry with the unit of work identified for the end of bracket entry, to see if the entry is related to the problem. Examine the entries with the same unit of work identifier in reverse timestamp sequence, starting with the entry before the end of bracket, then the entry before that, and so on, to try to identify all the relevant entries until you find the first entry for the initial cause of the problem.

    Some activity log entries are reraised exceptions received from lower level calls. The fact that these reraised exceptions occur suggests that these entries are not the source of the problem. Often, you are not interested in reraised exception entries in the log. Therefore, you may want to read the first few entries before the end of bracket and then quickly skim over the ones that have reraised exception

    Also, sometimes the runtime remaps the exception it receives from a lower call to another exception which is defined on its interface.

  3. Find the initial failure
    When you have found the first entry for the initial cause of the problem, you can take action to resolve the problem. Depending on the situation, you may also want to read a couple of entries before the initial failure's entry, just in case there is some additional data to help you diagnose the problem.