Issue
IMPAX Client is slow or hangs. This could be a continuous or intermittent issue.
The Oracle alert log file is a chronological log of important messages and errors that are encountered by the simple use of an Oracle database. Check the Oracle alert log for an error similar to the following:
Thread 1 cannot allocate new log, sequence 40857 Checkpoint not complete Current log# 3 seq# 40856 mem# 0: /usr/mvf/data/dbase/redo/redo03.log
Details
If "Checkpoint not complete" errors are happening repeatedly, the default number and/or size of the redo logs may be insufficient for the quantity of DML statements occurring at a site. If all available redo logs are active (except the current log) and are required for media recovery, there is no inactive redo log group and Oracle cannot switch to the next redo log; all DML activity stops and IMPAX slows down or hangs.
Solution
The database has a limited number of redo logs (three to five, depending on the size of the database). Resize the redo logs and/or add new ones. For example, instead of using 3 x 25MB redo logs, use 4 x 100MB.
Online redo logs cannot be resized; first, add the new redo logs and then drop the old ones.
Note:
For Oracle Data Guard sites, we strongly recommend that if you add or drop online redo log files from the primary database, you synchronize the changes on the standby database.
In the future, check for "Checkpoint not complete" messages in the alert log on a regular basis. On Solaris systems, the alert log is located at /usr/mvf/data/logs/oracle/bdump/alert_MVF.log. On Windows systems, the alert log is located at c:\mvf\data\logs\oracle\bdump\alert_MVF.log. On Solaris, you can run the following command:
cat /usr/mvf/data/logs/oracle/bdump/alert_MVF.log|grep 'Checkpoint not complete'
You can also check the status (active/inactive/current) of redo logs. Log into SQLPlus as the sys or dbadmin user.
Type
select * from v$sql;