httpd-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Brian Behlendorf <>
Subject Re: leaking fd's
Date Wed, 21 Jun 1995 04:04:38 GMT
On Tue, 20 Jun 1995, Rob Hartill wrote:
> which means that the parent SIGSEGV handler is getting in the way. Grrrr.

Oh yeah!  I also forgot to mention that recently we had another one of 
those run-away-with-the-CPU httpd's that wasn't leaving any sign by 
ktrace and burning up CPU in the kernel - however, doing a ktrace and 
sending the child a kill -HUP returned:

 22547 httpd   PSIG  SIGHUP caught handler=0x5ff0 mask=0x1400 code=0x0
 22547 httpd   CALL  sigprocmask(0x3,0)
 22547 httpd   PSIG  SIGSEGV caught handler=0x5e18 mask=0x0 code=0x0
 22547 httpd   RET   sigprocmask 5121/0x1401
 22547 httpd   CALL  gettimeofday(0xefbd572c,0)
 22547 httpd   RET   gettimeofday 0

... at which point it went back to running away with the CPU, not dying.  
A kill -HUp to the parent did clear it away though...

Looking forward to 0.7.3d :)


--=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=--  http://www.[hyperreal,organic].com/

View raw message