In reviewing the server.log on the NetBackup master server, there may be indications that Sybase occasionally exhausts the allocated threads. While symptoms may not appear due to connection retries, this does indicate a database unable to handle the load with the current threads allocated.
The server.log will display this error message below:
All threads were blocked when waiting to send or receive. A connection has been terminated. Increasing -gn may prevent this in the future.
server.log can be found in the following folder/directory:
Heavily utilized NetBackup master servers may run out of database threads under stress
The resolution for this issue (Etrack 3292302) is included in the following releases:
- NetBackup 7.5 Maintenance Release 7 (18.104.22.168)
- NetBackup 7.6 Maintenance Release 2 (22.214.171.124)
More information on these releases can be found in the Related Articles linked below.
If the server.log displays the error noted, please perform the following configuration changes in the server.conf file:
1. Verify the setting of the -ch switch for maximum RAM is set to 30% of the RAM available on the system. Adding more -gn threads to Sybase requires additional memory.
2. Add -gn in the server.conf startup to a value of 40 which will require a restart of NetBackup to take effect. The -gn switch adds more threads for use but requires more memory, this should not be set with the 7.5 default of -ch of 1GB. Default value without -gn specified is 20. Note: In NetBackup 7.6, the -gn should never be set as it acts to limit the thread settings.
3. If the message continues to be seen, contact Symantec technical support, referencing this document and Etrack 3387593 to obtain an Emergency Engineering Binary (EEB) for NetBackup 7.6 GA which will disable MAX_QUERY_TASKS. For NetBackup 126.96.36.199, please reference Etrack 3320755. After applying this EEB, a restart is not required.
4. Setting -gn higher than 40 should be done on a case by case basis. Check the emm.conf for thread settings that may be too high.