Error in lseek and ioctl

[laiju.c.babu]

Hi Team,

In our scenario we have 200+ linux servers. We are collecting the logs from all these  200+ linux servers to one hpunix machine. We are doing this activity using crontab and enabled passwordless ssh from  hpunix machine to all these linux servers so that we can do scp.

But for some of those machines we are not able  collect logs . I installed tusc and tracked the process, I got below errors from the tusc  output for some linux machines

lseek(3, 0, SEEK_CUR) .................................................................................................................. ERR#29 ESPIPE
ioctl(3, TCGETATTR, 0x40003960) ........................................................................................................ ERR#25 ENOTTY

I am able to do passwordless ssh to these linux servers and able to copy those logs manually using scp also. Could you please help me  out on this.

Regards

[Dennis Handly]

These are not the droids you seek.  I assume these errors are ignored and it continues.

What file is opened for FD 3?  Also what syscalls happens after them?

Are you using tusc on scp on HP-UX?

[laiju.c.babu]

Hi Dennis,

I am attaching the full system calls genrated for one server .

We have one script fo collecting logs from all other server (Linux+ Hpunix) We put this script in crontab . What i did  traced the process id of that script using tusc.

[Laurent Menase]

you should use -p -f options of tusc

[Dennis Handly]

>I am attaching the full system calls generated for one server.

This isn't full.  And as I said, these syscall errors at the start can be ignored.

>you should use -p -f options of tusc

Also -ea to trace execv call parms.

[laiju.c.babu]

Hi Dennis,

Thanks for the reply.

can i use the below syntax

tusc -p -f -ea <pid>

Regards

[Laurent Menase]

tusc -pfea -o /tmp/resultfile  command

or

tusc -pfea -o /tmp/resultfile  pid
or

tusc -pfea -o /tmp/resultfile  $$   &

scp command &

kill tusc when error is seen

-p -f -e -a is the same as -pfea

[Dennis Handly]

>can I use the below syntax

Yes.  I clump -ea together in case I want less output and want to remove both.

But for -fp, I always combine and use.

[laiju.c.babu]

I got the below output from the tusc

( Attached to process 650 ("sh -c     /eytools/eybin/get_linuxinfo.sh 1>/dev/null 2>&1") [32-bit] )
[650] waitpid(-1, 0x7fffe810, WUNTRACED) ............................................................................................... [sleeping]
[650] waitpid(-1, WIFEXITED(0), WUNTRACED) ............................................................................................. = 654
[650] sigvector(SIGCLD, 0x7fffe7e0, 0x7fffe7f0) ........................................................................................ = 0
[650] sigprocmask(SIG_SETMASK, 0x7fffe830, 0x7fffe850) ................................................................................. = 0
[650] sigprocmask(SIG_SETMASK, 0x7fffe850, 0x7fffe830) ................................................................................. = 0
[650] sigprocmask(SIG_SETMASK, 0x7fffe830, 0x7fffe850) ................................................................................. = 0
[650] sigprocmask(SIG_SETMASK, 0x7fffe850, 0x7fffe830) ................................................................................. = 0
[650] sigprocmask(SIG_SETMASK, 0x7fffe830, 0x7fffe850) ................................................................................. = 0
[650] sigprocmask(SIG_SETMASK, 0x7fffe850, 0x7fffe830) ................................................................................. = 0
[650] sigprocmask(SIG_SETMASK, 0x7fffe830, 0x7fffe850) ................................................................................. = 0
[650] sigprocmask(SIG_SETMASK, 0x7fffe850, 0x7fffe830) ................................................................................. = 0
[650] sigprocmask(SIG_SETMASK, 0x7fffe830, 0x7fffe850) ................................................................................. = 0
[650] sigprocmask(SIG_SETMASK, 0x7fffe850, 0x7fffe830) ................................................................................. = 0
[650] sigprocmask(SIG_SETMASK, 0x7fffe830, 0x7fffe850) ................................................................................. = 0
[650] sigprocmask(SIG_SETMASK, 0x7fffe850, 0x7fffe830) ................................................................................. = 0
[650] sigprocmask(SIG_SETMASK, 0x7fffe830, 0x7fffe850) ................................................................................. = 0
[650] sigprocmask(SIG_SETMASK, 0x7fffe850, 0x7fffe830) ................................................................................. = 0
[650] sigprocmask(SIG_SETMASK, 0x7fffe830, 0x7fffe850) ................................................................................. = 0
[650] sigprocmask(SIG_SETMASK, 0x7fffe850, 0x7fffe830) ................................................................................. = 0
[650] sigprocmask(SIG_SETMASK, 0x7fffe830, 0x7fffe850) ................................................................................. = 0
[650] sigprocmask(SIG_SETMASK, 0x7fffe850, 0x7fffe830) ................................................................................. = 0
[650] sigprocmask(SIG_SETMASK, 0x7fffe830, 0x7fffe850) ................................................................................. = 0
[650] sigprocmask(SIG_SETMASK, 0x7fffe850, 0x7fffe830) ................................................................................. = 0
[650] sigprocmask(SIG_SETMASK, 0x7fffe830, 0x7fffe850) ................................................................................. = 0
[650] sigprocmask(SIG_SETMASK, 0x7fffe850, 0x7fffe830) ................................................................................. = 0
[650] exit(0) .......................................................................................................................... WIFEXITED(0)

[Dennis Handly]

>I got the below output from the tusc

Nothing useful there.  How did you collect tusc info the first vs this time?