Nonstop Database, HiRDB Version 9 System Operation Guide

[Contents][Index][Back][Next]

20.22 Actions to take when an error occurs while a local buffer is being used to update a shared table (HiRDB parallel server configurations only)

Executor: HiRDB administrator

You must apply LOCK TABLE to use a local buffer to update a shared table. If the server process terminates abnormally when both of the conditions listed below are applicable, abort code Phb3008 is issued and the unit might terminate abnormally:

If all these conditions are satisfied and there is a page being updated when the server process terminates abnormally, transaction recovery processing might not be possible in some cases. If this happens, recovery processing is performed when the unit is restarted. The following table shows the HiRDB processing and the actions to take if an error occurs when a local buffer is being used to update a shared table (without the LOCK TABLE specification).

Table 20-26 HiRDB processing and actions to take if an error occurs when a local buffer is being used to update a shared table (without LOCK TABLE specification)

Cause of abnormal termination of server process HiRDB processing if an updated page exists when the server process terminates abnormally Action to be taken by the HiRDB administrator
PDSWAITTIME timeout Abort code Phb3008 is issued, and the unit might terminate abnormally. Restart the unit if it terminated abnormally.
PDCWAITTIME timeout
pdcancel command executed
Abort Abort that occurs due to detection of a HiRDB conflict.
Other SIGSEGV, SIGBUS, receiving of external signal, exit, and other unexpected errors

Note
HiRDB does not terminate abnormally when a server process terminates abnormally due to a PDSWATCHTIME timeout, even if an updated page exists.