https://community.hpe.com/t5/operating-system-hp-ux/system-call-trace/m-p/3625269#M236515 You need to use the -c option of tusc to get a count of each system call. You need to get some idea of the fraction of total time each system call is using. This behavior may be perfectly normal but tusc by itself is not that useful . You really need to combine it with Glance (or at least sar output) so that this can be put into context. Is your timeslice set to 1? Tue, 13 Sep 2005 21:06:14 GMT A. Clay Stephenson 2005-09-13T21:06:14Z https://community.hpe.com/t5/operating-system-hp-ux/system-call-trace/m-p/3625268#M236514 I found process have much of cpu usage.<BR />so i used tusc tool. but I don't know how can I handle it. could you check this and is there any ways to check tusc more efficiently?<BR /><BR />this is tusc results.<BR /><BR />[15210] nanosleep(0x4f0fb398, NULL) ....................................... = 0<BR />[15210] nanosleep(0x4f0fb398, NULL) ....................................... = 0<BR />[15210] nanosleep(0x4f0fb398, NULL) ....................................... = 0<BR />[15210] nanosleep(0x4f0fb398, NULL) ....................................... = 0<BR />[15210] nanosleep(0x4f0fb398, NULL) ....................................... = 0<BR />[15210] poll(0x4edf5b80, 1, 4000) ......................................... = 0<BR />[15210] sigprocmask(SIG_BLOCK, 0x77fbe018, 0x4edf5be4) .................... = 0<BR />[15210] sigprocmask(SIG_SETMASK, 0x4edf5be4, NULL) ........................ = 0<BR />[15210] sigprocmask(SIG_BLOCK, 0x77fbe018, 0x4edf5b9c) .................... = 0<BR />[15210] sigprocmask(SIG_SETMASK, 0x4edf5b9c, NULL) ........................ = 0<BR />[15210] nanosleep(0x4ea6e9d8, NULL) ....................................... = 0<BR />[15210] nanosleep(0x4f0fb398, NULL) ....................................... = 0<BR />[15210] open("/oracle10g/bizmax_HP10G/sysman/recv", O_RDONLY|0x200000, 07) = 79<BR />[15210] fstat(79, 0x4ea6eb08) ............................................. = 0<BR />[15210] fcntl(79, F_SETFD, 1) ............................................. = 0<BR />[15210] getdents(79, 0x1b5f530, 8192) ..................................... = 72<BR />[15210] getdents(79, 0x1b5f530, 8192) ..................................... = 0<BR />[15210] close(79) ......................................................... = 0<BR />[15210] poll(0x4eff9a40, 1, 4000) ......................................... = 0<BR />[15210] sigprocmask(SIG_BLOCK, 0x77fbe018, 0x4eff9aa4) .................... = 0<BR />[15210] sigprocmask(SIG_SETMASK, 0x4eff9aa4, NULL) ........................ = 0<BR />[15210] sigprocmask(SIG_BLOCK, 0x77fbe018, 0x4eff9a5c) .................... = 0<BR />[15210] sigprocmask(SIG_SETMASK, 0x4eff9a5c, NULL) ........................ = 0<BR />[15210] nanosleep(0x4f0fb398, NULL) ....................................... = 0<BR />[15210] nanosleep(0x4f0fb398, NULL) ....................................... = 0<BR />[15210] ksleep(PTH_CONDVAR_OBJECT, 0x81450, 0x81458, 0x6af4e4a0) .......... = -ETIMEDOUT<BR />[15210] gettimeofday(0x6af4e4a8, NULL) .................................... = 0<BR />[15210] clock_gettime(CLOCK_REALTIME, 0x6af4e540) ......................... = 0<BR />[15210] nanosleep(0x4f0fb398, NULL) ....................................... = 0<BR />[15210] nanosleep(0x4f0fb398, NULL) ....................................... = 0<BR />[15210] nanosleep(0x4e9eda58, NULL) ....................................... = 0<BR />[15210] nanosleep(0x4f0fb398, NULL) ....................................... = 0<BR />[15210] nanosleep(0x4f0fb398, NULL) ....................................... = 0<BR />[15210] nanosleep(0x4f0fb398, NULL) ....................................... = 0<BR />[15210] nanosleep(0x4f0fb398, NULL) ....................................... = 0<BR />[15210] nanosleep(0x4f0fb398, NULL) ....................................... = 0<BR />[15210] nanosleep(0x4f0fb398, NULL) ....................................... = 0<BR />[15210] nanosleep(0x4f0fb398, NULL) ....................................... = 0<BR />[15210] nanosleep(0x4f0fb398, NULL) ....................................... = 0<BR />[15210] nanosleep(0x4f0fb398, NULL) ....................................... = 0<BR />[15210] nanosleep(0x4f0fb398, NULL) ....................................... = 0<BR />[15210] nanosleep(0x4f0fb398, NULL) ....................................... = 0<BR />[15210] nanosleep(0x4f0fb398, NULL) ....................................... = 0<BR />[15210] nanosleep(0x4f0fb398, NULL) ....................................... = 0<BR />[15210] nanosleep(0x4f0fb398, NULL) ....................................... = 0<BR />[15210] nanosleep(0x4f0fb398, NULL) ....................................... = 0<BR />[15210] ksleep(PTH_CONDVAR_OBJECT, 0x81450, 0x81458, 0x6af4e4a0) .......... = -ETIMEDOUT<BR /> Tue, 13 Sep 2005 20:41:44 GMT https://community.hpe.com/t5/operating-system-hp-ux/system-call-trace/m-p/3625268#M236514 ???_185 2005-09-13T20:41:44Z https://community.hpe.com/t5/operating-system-hp-ux/system-call-trace/m-p/3625269#M236515 You need to use the -c option of tusc to get a count of each system call. You need to get some idea of the fraction of total time each system call is using. This behavior may be perfectly normal but tusc by itself is not that useful . You really need to combine it with Glance (or at least sar output) so that this can be put into context. Is your timeslice set to 1? Tue, 13 Sep 2005 21:06:14 GMT https://community.hpe.com/t5/operating-system-hp-ux/system-call-trace/m-p/3625269#M236515 A. Clay Stephenson 2005-09-13T21:06:14Z