|
On Thu, 2003-06-19 at 00:39, Roberto Nibali wrote:
> Robert Lazzurs wrote:
> > I have setup a test network with a few app servers and a lvs machine
> > with one client. I seems to have issues topping 5000 connections per
> > second.
>
> Which count? ipvsadm -L -n or ab? What exactly do both tell you at the
> point of saturation? What kind of NICs do you use? Could you check their
> speed setting with either mii-tool or ethtool, please?
Supports auto-negotiation: Yes
Speed: 100Mb/s
Duplex: Full
The count was from ab however I checked when doing the same tests and it
was 500 per real server.
The count from ab was as below
Requests per second: 2948.98 [#/sec] (mean)
and I was getting this from ipvsadm
TCP 192.168.55.5:http rr
-> 192.168.55.1:http Route 1 500 14699
-> 192.168.55.3:http Route 1 500 15404
> > I also tried to patch httperf as per their suggestion but when I did
> > this it just sat and ate cpu time.
>
> ;).
Should the patch work or should I be doing something else?
> > I have tried the above tests (ab) on both the real server and the
> > virtual server and I get about the same results. It is not a bandwidth
>
> I do not understand this, sorry. What is the network flow for the two
> mentioned test conducts above?
>
> test ----> LVS ----> RS
> test --------------> RS
Yea, that is what I have been doing, sorry for not explaining it
clearly.
> > issue as all of the machines are on a 100Mbit switch and it is not
> > coming close to maxing out the bandwidth of any of the machines.
>
> It can also be a pps (packets per seconds) issue.
How would I be able to check if this is the case and how would I be able
to solve it?
> > I have also checked that it is not the test client machine by adding
> > another client machine in as well and if I add the connections per
> > second of both of them I get about the same results.
>
> From short thinking I see the problems in two endroits:
>
> o thttpd is not able to handle more connections
Well it was able to take more than apache and I tried setting that to
take the most connections it could. Do you have any better suggestions
on software I should be using client side, even another protocol.
> o network congestion in regard of pps
Well it is all on it's own switch so I doubt that is the issue.
> > If anyone has any suggestions on how I should proceed then that would be
> > most excellent, use kernel 2.5? change the web server? what ever you
> > think it could be.
>
> Could you try with apache as well please and could you provide me with
> the numbers I requested in the beginning of this email? I does
> definitely not look like an LVS problem as with your setup you would be
> able to handle 5000 pps with a 3 Minute persistency even.
I know it should be able to handle more but it appears there is
something wrong with my tests.
However I do get this from both of the app servers
TCP: time wait bucket table overflow
I tried google and usenet however I could not find anything useful
Thanks again for your help, take care - RL
--
MSN:lazzurs@xxxxxxxxxxxxxx |"All that is etched in stone
Yahoo:admroblaz AIM:admroblaz |is truly only scribbled in
ICQ:66324927 |sand" - RL
Jabber:admroblaz@xxxxxxxxxx |Join Eff http://www.eff.org
e-mail:lazzurs@xxxxxxxxxxxxxxxxx|Take care all - Rob Laz
|
