Quoting <strong>Alan DeKok &lt;<a href="mailto:aland@deployingradius.com">aland@deployingradius.com</a>&gt;</strong>: <blockquote style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">Konstantin Chekushin wrote:<br /> <sigh><br /> &gt; I've started<br /> &gt; ./radius -xxx -f<br /> <br />   "radius"?  Not "radiusd"?  And why not "radiusd -X" as documented<br /> *everywhere* ?<br /> </sigh></blockquote>-X = -sfxx -l stdout . But, I didn't wanted single mode for my loading test. So I used "./radiusd -xxx -f" (the letter was missed in a previous mail)<br />And I've used -xxx to make my my own debug3 visible with timestamp (without changing your log.c). <br />In fact, your logging architecture logic is rather strange. But this is another topic.<br /><br /><blockquote style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;"><sigh><br /> &gt; I've prepared huge file with auth requests for sending:<br /> <br />   Which no one asked for, and isn't necessary.<br /> <br /> &gt; my module process requests too slow (I will discover this problem in few<br /> &gt; days, this is mysql-db-cluster issues), but this test causes radiusd<br /> &gt; segmentation fault. I've grabbed the core and look what it shows:<br /> <br />   This should be fixed in 2.1.10.<br /> </sigh></blockquote><br />Good. Thanks for the answer.