[httperf] Bad file descriptor in select

Martin Arlitt arlitt at granite.hpl.hp.com
Thu Feb 15 09:27:01 PST 2007


hi Jordi

I should have asked what the session duration was, that will affect the
connection durations.

if the problem is not TIME_WAIT, then another possibility (and this is a
guess, because I haven't tested it) is its a mismatch between FD_SETSIZE
and the kernel setting for the local ip port range:
/proc/sys/net/ipv4/ip_local_port_range

what does
$ cat /proc/sys/net/ipv4/ip_local_port_range
return?

if it's not related to exhausting available file descriptors, then I'm not
sure what else to check.  is the tomcat server reseting any of the
open/active connections?

Martin

> >hi Jordi
> >
> >if it's repeatable, that's good.
> >
> >are you sure you are not running out of file descriptors? how many file
> >descriptors does httperf have access to? if you run httperf with the
> >--verbose option that should print a line like
> >httperf: maximum number of open descriptors = 1024
> >
> >
> httperf: maximum number of open descriptors = 65000
>
> >since you stated the test runs for 2-3 minutes, I'm not sure this is the
> >problem (I'd expect it to happen a bit sooner), but it's probably worth
> >checking.
> >
> >if you are not sure if it is running out of file descriptors, one way to
> >check is the following; re-run your experiment with a high request rate
> >until the error occurs. as soon as the error occurs, check how many
> >connections are in the TIME_WAIT state.
> >depending on which OS you are using, you should be able to run the
> >following command:
> >
> >netstat -an | grep TIME_WAIT | wc -l
> >
> >if that returns a number about as large as the number of file descriptors
> >available to httperf, then that is likely the cause.
> >
> >
> This command returns 0
>
> >what OS are you running on?
> >
> >
> I run httperf on a Debian distribution with kernel 2.6.16
>
> Do you think the error could be related with the FD_SETSIZE value?
>
> Jordi
>
> >thanks
> >
> >Martin
> >
> >On Thu, 15 Feb 2007, Jordi Guitart Fernandez wrote:
> >
> >
> >
> >>Hello,
> >>
> >>I'm using httperf to issue secure requests to a Tomcat server. I run
> >>httperf with these parameters:
> >>
> >>httperf --hog --timeout=10 --client=0/1 --server=pcbosch.ac.upc.edu
> >>--port=8443 --uri=/ --rate=50 --send-buffer=4096 --recv-buffer=16384
> >>--ssl --wsesslog=100000000,0.000,rubis.txt
> >>
> >>I face with the "bad file descriptor" error when using high rates (i.e.
> >> >= 40), and it occurs after 2 or 3 minuts of experiment.
> >>
> >>Jordi
> >>
> >>Martin Arlitt escribió:
> >>
> >>
> >>
> >>>hi Jordi
> >>>
> >>>I haven't experienced this error before, so this is just a guess.
> >>>
> >>>it's probably the following error (errno.h):
> >>>#define EBADFD          77      /* File descriptor in bad state */
> >>>
> >>>I believe this could happen if httperf tries to read from a file
> >>>descriptor that has been closed, for example.  if you can provide more
> >>>details on what you were doing when this error occurred, perhaps it will
> >>>shed some more light on the cause.
> >>>
> >>>thanks
> >>>
> >>>Martin
> >>>
> >>>On Thu, 15 Feb 2007, Jordi Guitart Fernandez wrote:
> >>>
> >>>
> >>>
> >>>
> >>>
> >>>>Hello,
> >>>>
> >>>>What can cause httperf to return the following error?
> >>>>
> >>>>"httperf.core_loop: select failed: Bad file descriptor"
> >>>>
> >>>>Thx
> >>>>_______________________________________________
> >>>>httperf mailing list
> >>>>httperf at linux.hpl.hp.com
> >>>>http://www.hpl.hp.com/hosted/linux/mail-archives/httperf/
> >>>>
> >>>>
> >>>>
> >>>>
> >>>>
> >>>_______________________________________________
> >>>httperf mailing list
> >>>httperf at linux.hpl.hp.com
> >>>http://www.hpl.hp.com/hosted/linux/mail-archives/httperf/
> >>>
> >>>
> >>>
> >>>
> >>_______________________________________________
> >>httperf mailing list
> >>httperf at linux.hpl.hp.com
> >>http://www.hpl.hp.com/hosted/linux/mail-archives/httperf/
> >>
> >>
> >>
>
>



More information about the httperf mailing list