General
cancel
Showing results for 
Search instead for 
Did you mean: 

Pid ... killed due to trashed stack, Oracle 10g

Andy Meltzer
Occasional Visitor

Pid ... killed due to trashed stack, Oracle 10g

We are getting the error "Pid 26693 killed due to trashed stack. Pid 26693 was killed due to failure in writing the signal context" when we try to run our application on HP UX 11.11 ( Oracle 10g ). But the same execution is working fine on HPUX 11.11 ( Oracle 9i ). Does anyone know of a solution to this problem?
4 REPLIES
Dennis Handly
Acclaimed Contributor

Re: Pid ... killed due to trashed stack, Oracle 10g

You could have a (thread?) stack overflow, where there is no room to write the context.

Do you have a stack trace?
Andy Meltzer
Occasional Visitor

Re: Pid ... killed due to trashed stack, Oracle 10g

The stack isn't very deep, and it is dies in libclntsh.sl. Is there anything specific I should be looking for?
Dennis Handly
Acclaimed Contributor

Re: Pid ... killed due to trashed stack, Oracle 10g

What does this show in gdb:
bt
info reg
p /x $sp_save = $sp
frame XX # the largest frame number in the stack trace
p $sp_save - $sp
frame 0
disas $pc-4*8 $pc+4*4

Andy Meltzer
Occasional Visitor

Re: Pid ... killed due to trashed stack, Oracle 10g

Dennis,

Thanks for the help. The stack pointed us in the right direction, but the problem turned out to be with too many pthreads being allocated.
We've upped a parameter in the program causing the problem.