Question & Answer
Question
When Data Protection for Domino process is terminated improperly, it may cause the Domino server to crash. Is there a way to safely stop the domdsmc/domdsm process?
Cause
The current Domino design can invoke a Domino Server crash when a client process is terminated abnormally. For example, terminating Data Protection for Domino with a control-C or a kill command would abruptly stop all processing, however the processing might not be terminated cleanly and result in errors such as “Unknown OS error” or “ACD5702E A Domino API error has occurred”. In the worst case, the Domino server can crash with a fatal error.
Abnormal termination of the Data Protection for Domino can result from the following actions:
Examples on Windows:
- end the domdsmc.exe process from Windows Task Manager
- close the domdsmc.exe command window in the middle of the backup/restore
- close the Data Protection for Domino GUI (domdsm.exe) in the middle of the backup/restore
- kill -9 <domdsmc pid>
- ctrl-C to close the domdsmc command window in the middle of the backup/restore
Answer
To stop the Data Protection for Domino client when it is in the middle of processing, this needs to be performed from the Tivoli Storage Manager Server by cancelling the session.
The Tivoli Storage Manager server administrator can terminate an individual database backup by cancelling the Data Protection for Domino session from the Tivoli Storage Manager server. However, the Data Protection product is designed to retries all errors, including communications failures caused by a Tivoli Storage Manager server disconnecting the session. If many databases were being backed up, each database communications failure would be retried, making it very difficult to totally cancel the backup processing.
To accommodate the termination of the backup, the Data Protection for Domino communications processing of the Tivoli Storage Manager administrator cancel session was changed in version 6.3. When a Tivoli Storage Manager administrator cancel session request is received, Data Protection for Domino version 6.3 (or greater) will treat this as a request to cancel all processing. The processing is ended at a clean break point, rather than at an arbitrary point, and Data Protection for Domino terminates with the message “ACD0592E The Data Protection TCP/IP session with the Tivoli Storage Manager server was canceled.”
If there are multiple Data Protection for Domino sessions, it is sufficient to cancel one of the Data Protection for Domino sessions. When using Lanfree communication. You must cancel the session that has the session stage RecW .
An example of using the Tivoli Storage Manager administrator cancel session command to terminate Data Protection for Domino processing is shown below.
Tivoli Storage Manager server administrator issues a query session command to determine the Data Protection for Domino session number and then cancels that session.
- tsm: TSMSSERVER1>cancel sess 1761
ANR0490I Canceling session 1761 for node NOTES2 (TDP Domino AIX64) .
When Data Protection for Domino detects the cancel session request, the current operation is terminated cleanly and Data Protection for Domino exits without processing any additional databases. The following messages will be written to the Data Protection to Domino log as a result of the cancel session request.
- ACD5450E A failure occurred on Tivoli Storage Manager server session number (1), rc = 118
- ANS1017E (RC-50) Session rejected: TCP/IP connection failure
- Backup of events4.ntf failed.
- ANS1017E (RC-50) Session rejected: TCP/IP connection failure
- ACD0590E A Data Protection communication error with the Tivoli Storage Manager server has occurred.
- ACD0592E The Data Protection TCP/IP session with the Tivoli Storage Manager server was canceled.
Product Synonym
TSM
Was this topic helpful?
Document Information
Modified date:
17 June 2018
UID
swg21663277