[Tor-BSD] Recognizing Randomness Exhaustion
Pete Wright
pete at nomadlogic.org
Tue Jan 6 01:26:15 EST 2015
On 1/1/15 9:23 PM, Libertas wrote:
> However, we're now finding that a profusion of gettimeofday() syscalls
> may be the issue. It was independently discovered by the operator of
> IPredator, the highest-bandwidth Tor relay:
>
> https://ipredator.se/guide/torserver#performance
>
> My 800 KB/s exit node had up to 7,000 gettimeofday() calls a second,
> along with hundreds of clock_gettime() calls.
i am seeing quite a few clock_gettime requests on a freebsd-10.1 system:
pete at tor-tracks:~ % sudo dtrace -n 'syscall::clock_gettime:entry {
@[execname] = count(); }'
dtrace: description 'syscall::clock_gettime:entry ' matched 1 probe
^C
syslogd 1
sendmail 3
ntpd 7
sshd 8
dtrace 26
tor 25274
dtruss shows similar numbers:
pete at tor-tracks:~ % sudo dtruss -n tor -c -t clock_gettime
<snip>
CALL COUNT
close 1
sendto 2
recvfrom 4
getpid 18
kevent 135
write 154
read 412
gettimeofday 447
clock_gettime 2837
>
> Because IPredator runs Linux, he used vsyscalls to speed things up.
> We'll probably need to find something more creative, like using our time
> caching more.
>
> We're working on it with this ticket:
>
> https://trac.torproject.org/projects/tor/ticket/14056
>
thanks for the link i'll keep an eye on that ticket.
cheers,
-pete
--
Pete Wright
pete at nomadlogic.org
twitter => @nomadlogicLA
More information about the Tor-BSD
mailing list