Always send the whole error message. If this is all you see, then something is misbehaving and you'll have to catch the message in a debugger.<br><br><div class="gmail_quote">On Mon, Apr 9, 2012 at 20:10, Mohammad Mirzadeh <span dir="ltr"><<a href="mailto:mirzadeh@gmail.com">mirzadeh@gmail.com</a>></span> wrote:<br>
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div dir="ltr">Hi,<div><br></div><div>I have this strange behavior that sometimes when I run the code on certain number of processors and for a certain problem size, I get a message like:</div>
<div><br></div><div>"key # not found!"</div>
<div><br></div><div>where # is just some number (like 9087 but changes every time I run it). Is there something wrong with my MPI? is this even an MPI message or something else? Sorry my question is this broad, but I have no clue where this message is popping from! PETSc does not produce any error message when this happens and the code is valgrind clean.</div>
<div><br></div><div>I'd appreciate any input.</div><div>Thanks</div></div>
</blockquote></div><br>