The executable for a particular SCM cannot be found
Explanation: The SourceIQ server is configured with default locations for executables required to execute SCM commands. Some of these tools (e.g., CVS) are delivered with the SourceIQ installation, and so should not manifest this issue. Others (e.g., AccuRev, ClearCase) require a path to an executable outside the scope of the SCPM deployment directory, and therefore may need to be adjusted on a particular platform.
Fix: Two sets of changes are required. The first will ensure that newly created SCMs will have the correct executable path. The second is required for any SCMs created with the previously incorrectly configured path.
In order to change the default executable paths for NEW SCMs, open the SourceIQ database, open the REF_REPOSITORY_TYPES table, and modify the value of the REPOSITORY_CLIENT column of the row with appropriate REPOSITORY_TYPE.
For previously created SCMs, open the SCM_CONNECTIONS table and modify the value of the CLIENTPATH column for the affected SCMs.
Hard sandbox expiration before completion
Explanation: This will only occur if the server is in the hard sandbox scheduling mode. The task was not able to complete before the server needed to stop acquitions in accordance with the hard sandbox parameters.
Fix: Nothing. When the server is allowed to resume acquistions it will pick up the task and continue with the acquisition.
Transaction failed due to timeout...
Explanation: A database transaction took longer to complete than configured for a specified operation. The error message should indicate the phase in which the timeout occurred and the duration of the operation than failed.
Fix: There are several configuration settings that can be adjusted to address this issue.
If the error indicates that the problem occurred in the Database Reconciliation phase (the most common instance), open the Conduits dialog and modify the Conduit Reconciliation Timeout (secs) value - Conduit Timeout (secs) prior to version 4.5 - for the Conduit that timed out. If most or all of the Conduits on a Server manifest this issue, you can modify the Default Reconciliation Timeout (secs) value in the Servers dialog. If the environment has multiple Servers (a Distributed Acquisition environment), modify the setting for the Server(s) hosting the problematic Conduits.
If the error indicates that the problem occurred in any other phase, this is generally a problem with a specific (usually large) revision. In this instance, you need to modify the Per-Revision Timeout (secs) value for the specific Conduit in the Conduits dialog. There is no Server-wide setting for this value.
In general, you should adjust the setting you are modifying to some value in seconds greater than that indicated by the error message.
If these modifications do not correct the problem, please contact support@sourceiq.com.
StackOverflowException
NOTE - As opposed to most SourceIQ exceptions, a StackOverflowException will not be reported as the cause of a failure in the Tasks dialog. Rather, the Task will report a failure, but the cause will be empty. This error will only be manifested in the JBoss server.lop file. This is because the error bypasses the normal Java exception handling mechanisms. SourceIQ is currently investigating options to capture this error in the Task failure reports, as well as looking at alternative strategies to avoid the issue altogether.
Explanation: Some acquisition processing has caused the internal JBoss stack limitations to exceed their configured boundaries.
Fix: The JBoss stack configuration needs to be adjusted. See http://open.sourceiq.com:18080/jamwiki/en/General_Server_Setup#JBoss_Configuration for details on making this adjustment.