Troubleshooting the Xinet database

Xinet is complex, and as such there are different ways that things can go wrong. This chapter addresses those issues and tools with which to gather meaningful information about problems and solve them.

The issues discussed include:

•	The mysqld daemon can’t be restarted.

•	The database tables can be corrupt.

•	Events cannot be added to the database.

•	Particular events can stop being added correctly to the database.

•	The contents of the file system may not necessarily match what’s in the database (usually only in one area of the database).

•	Searches can be slow.

•	Search results can be incorrect.

•	Xinet programs start consulting the file system rather than the database.

•	Particular events can be processed very slowly by the database (or never get processed).

•	Problems with Office Web previews.

In order to address these issues (and others like them) it helps to have a general idea of how to get meaningful information from a “broken” database. The next section will deal with some general things to look for when you begin troubleshooting. If you need an introduction to the binaries that do most of the work in Xinet, see Database synchronization and About the dblogd(8) daemon.

The following topics are provided in this section:

•	When the file system and the database don't match

•	Other important questions to ask when troubleshooting database issues

•	Further steps for troubleshooting

•	The venture.log and venturelog.old files

•	The fpod(1M) vlog file

•	The SERVER.err file

•	The at_log file

•	Logs from Xinet queues

•	Problems with Office Web previews