| Summary | How to address performance degradation of an Orbix process? |
|---|---|
| Article Number | 27151 |
| Environment | Product: Orbix Version: 3.x OS: all |
| Question/Problem Description | An Orbix application observed performance degradation. Event lists are getting queued up and crossing MAX-threshold limits , there by causing severe damage . The naming service thread dump is as following: % psg orbix 13409 13408 0 Oct 05 ? 1:45 /usr/java/bin/java -Xbootclasspath/ /14: lwp_cond_wait(0x000363F0, 0x000363D8, 0xB3D7FA60, 0) Err#62 ETIME /14: condvar type: USYNC_THREAD /14: mutex type: USYNC_THREAD /14: timeout: 0.999997700 sec /22: pollsys(0x00000000, 0, 0xB357FA90, 0x00000000) = 0 /22: timeout: 0.050000000 sec /22: pollsys(0x00000000, 0, 0xB357FA90, 0x00000000) = 0 /22: timeout: 0.050000000 sec /22: pollsys(0x00000000, 0, 0xB357FA90, 0x00000000) = 0 /22: timeout: 0.050000000 sec /22: pollsys(0x00000000, 0, 0xB357FA90, 0x00000000) = 0 /22: timeout: 0.050000000 sec /44: read(10, 0xB247EF88, 4) (sleeping...) /33: lwp_cond_wait(0x0026A2B0, 0x0026A298, 0x00000000, 0) (sleeping...) /33: condvar type: USYNC_THREAD /33: mutex type: USYNC_THREAD % truss -c -p 13409 syscall seconds calls errors lwp_cond_wait .000 11 11 pollsys .005 184 -------- ------ ---- sys totals: .005 195 11 usr time: .001 elapsed: 10.990 |
| Clarifying Information | |
| Error Message | 3409 13408 0 Oct 05 ? 1:45 /usr/java/bin/java -Xbootclasspath/ /14: lwp_cond_wait(0x000363F0, 0x000363D8, 0xB3D7FA60, 0) Err#62 ETIME /14: condvar type: USYNC_THREAD /14: mutex type: USYNC_THREAD /14: timeout: 0.999997700 sec /22: pollsys(0x00000000, 0, 0xB357FA90, 0x00000000) = 0 /22: timeout: 0.050000000 sec /22: pollsys(0x00000000, 0, 0xB357FA90, 0x00000000) = 0 /22: timeout: 0.050000000 sec /22: pollsys(0x00000000, 0, 0xB357FA90, 0x00000000) = 0 /22: timeout: 0.050000000 sec /22: pollsys(0x00000000, 0, 0xB357FA90, 0x00000000) = 0 /22: timeout: 0.050000000 sec /44: read(10, 0xB247EF88, 4) (sleeping...) /33: lwp_cond_wait(0x0026A2B0, 0x0026A298, 0x00000000, 0) (sleeping...) /33: condvar type: USYNC_THREAD /33: mutex type: USYNC_THREAD % truss -c -p 13409 syscall seconds calls errors lwp_cond_wait .000 11 11 pollsys .005 184 -------- ------ ---- sys totals: .005 195 11 usr time: .001 elapsed: 10.990 |
| Defect/Enhancement Number | |
| Cause | This is due to memory exhaustion. |
| Resolution | The resolution to the performance degradation is increasing the memory allocated to the process In question. In case it is a java process, like the naming service, This can be done by adding the following parameters after the java command: -Xms 1024m -Xmx 1024m |
| Workaround | |
| Notes | |
| Attachment |
| Created date: | 11 December 2011 |
|---|---|
| Last Modified: | 14 February 2013 |
| Last Published: | 11 December 2011 |
| First Published date: | 11 December 2011 |
#KnowledgeDocs
#Orbix
