On 07/09/2009, Oliver Erlewein [DATACOM] <[email protected]> wrote: > Hi, > > Could be that I just reposted my question by accident. Sorry! > > I don't think the throttling code is too specific. The perl script does 100% > the same thing and is fully rate limited. I'm totally stumped because I'd say > everything uses the same kernel modules no matter what language it is.
Agreed. > Is there any way I can get more detailed information out of JMeter/Java on > what's happening (or not)? You could try using a protocol analyser such as Wireshark - this would show if the packets look the same. I don't know how to check what kernel modules are being used. Have you tried posting the question on a forum dedicated to the OS you are using? > Cheers > > Oliver > > > -----Original Message----- > From: sebb [mailto:[email protected]] > Sent: Monday, 7 September 2009 10:13 a.m. > To: JMeter Users List > Subject: Re: JMeter and rate limiting with TC & IPtables > > On 03/09/2009, Oliver Erlewein [DATACOM] <[email protected]> > wrote: > > Hi all, > > > > I throttled the bandwidth on a Ubuntu box for a test. The script looks > something like this (snippets): > > > > tc qdisc add dev eth0 handle ffff: ingress > > tc filter add dev eth0 parent ffff: protocol ip prio 50 u32 match > ip src 172.xx.xx.0/24 \ > > police rate 34Kbits burst 5000 mtu 5000 drop flowid :1 > > tc qdisc add dev eth0 root tbf rate 34Kbits latency 25ms burst > 5000 > > > > The target systems are all on 172.xx.xx.0 . There's a perl script that > loads pages that get's throttled just fine. When I run JMeter with an > equivalent script it does not get throttled. > > > > Has anyone tried this or experimented with this? If so, does JMeter/Java > circumvent the normal TCP stack or something? > > JMeter uses standard Java sockets. > Seems rather unlikely that Java would use different system code from > Perl, so perhaps the filter is too specific. > > > Oh and I am aware JMeter can throttle too but in my situation I can't use > that function. > > > > Cheers > > Oliver > > > > --------------------------------------------------------------------- > > To unsubscribe, e-mail: [email protected] > > For additional commands, e-mail: [email protected] > > > > > > --------------------------------------------------------------------- > To unsubscribe, e-mail: [email protected] > For additional commands, e-mail: [email protected] > > > --------------------------------------------------------------------- > To unsubscribe, e-mail: [email protected] > For additional commands, e-mail: [email protected] > > --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
