TM1 10.2.2. FP7 - Managing TM1 server shutdown due to network delays that cause transaction log update failures
As of TM1 10.2.2 Fix Pack 7 (FP7), you can prevent automatic TM1 server shutdowns when network issues block access to transaction logs. This feature is enabled through the use of three new tm1s.cfg parameters.
Planning Analytics users can manage server shutdown behavior when the system encounters network delays during transaction log updates. Previously, when a network failure occurred, TM1 would automatically self-terminate. The new FileRetry.FileSpec, FileRetry.Count, and FileRetry.Delay tm1s.cfg parameters allow users to prevent the automatic server shutdown. Enabling this feature defers the shutdown, while the server attempts to reestablish access to the transaction logs.
- Specifies the directory path(s) of the affected log files.
- String value
- Default value is <empty>
- Specifies the number of retry attempts.
- Integer value
- Default value: 5
- Specifies the delay (in milliseconds) between retry attempts.
- Integer value
- Default value: 2000
All three parameters are "dynamic" and are refreshed in the background. To enable the feature, specify the directory path of the affected files in the FileRetry.FileSpec parameter. Use a semi-colon delimiter to specify multiple paths.
All files in the "c:\production\model" directory and below have the retry logic applied on a write failure.
All files in the "c:\production\model\Logs" directory and below have the retry logic applied on a write failure.
All files in the "c:\production\model\Logs" directory and the network share of \\network.ibm.com\production\Logs and below have the retry logic applied on a write failure.
Note: Messages are written to the server log at the WARN level when this logic is enabled.
- "Error writing to <FILENAME>, retry in progress." is written when a write operation failed and a retry attempt is in-progress.
- "Error writing to <FILENAME>, retry attempt failed." is written when the retry attempt failed.