On Tue, 12 Jun 2001, Joseph Mack wrote:
> > > In the meantime, how can I start 256 clients, rather than the 1024
>
> > So, to get 256 robots, you can set peak_req_rate to
> > peak_req_rate = client_side.max_agent_load * 256;
>
> 256 is replacing count(client_side.hosts) which in my example =1.
It is not. In my formula, I have client_side.max_agent_load as a
multiplier. Your file has client_side.max_host_load. There is a three
orders of magnitude difference. :)
> polyclt exits straight away saying something like "too many
> clients".
Right. Your "client_side.max_host_load * 256" is 179,200 req/sec load.
The addressing scheme does not have enough IP address to generate that
kind of load. My formula should work though.
> If I put 0.5 instead of 256, I get the client to fire up.
With "client_side.max_host_load * 0.5" you get (700req/sec * 0.5 =
350req/sec) which is (350req/sec / 0.7req/sec = 500 robots).
> > Make sure you re-configure and re-compile Polygraph when you change FD
> > limits.
>
> No change. The code has code like
>
> Min(FD_SETSIZE, rlimit) (eg in PolyApp.cc) which picks up 1024. I
> haven't found where it picks up rlimit yet (there are no likely
> suspects in the header files in PolyApp.c).
FD_SETSIZE is an OS constant that Polygraph obeys on Linux. Rlimit is
obtained using getrlimit(2) system call, I think. Perhaps the
following URLs will help you to configure your Linux box.
http://www.squid-cache.org/Doc/FAQ/FAQ-11.html#ss11.4
http://squid.sourceforge.net/hno/linux-lfd.html
> >As virtually everything else, these addresses come from your
> >configuration file:
> > server_side.addr_mask = '192.168.1.0:8080';
> > vipMap.addresses = serverAddrs(asWebAxe4, TheBench);
> > vipMap.names = addresses;
> > ClientR.origins = vipMap.names;
> >
> >Since asWebAxe4 is returning aliases based on addr_mask, you
> >probably want to replace the second line with
> > vipMap.addresses = server_side.hosts;
>
> I get "server_side.hosts or its components not declared"
> when I try to run the client.
It's TheBench.server_side.hosts. I omitted a few prefixes to save
some typing. Simply search for those field names in your file, most of
the names are unique so it is easy to find the object they belong to.
> I still don't see how I get the servers listening on 192.168.128.1:8080
> when I put
>
> server_side.addr_mask = '192.168.1.0:8080';
>
> The servers listen on the same address if I have the address_mask
> 192.168.255.255 or 192.168.0.0. It appears that the
> server_side.addr_mask isn't affecting the address the server is
> listening to.
You are using the serverAddrs(asWebAxe4, TheBench) call to set
OriginS.addresses (directly or indirectly). AsWebAxe4 uses
server_side.addr_mask to build addresses. However, it ignores last two
octets of the mask because, based on WebAxe rules, the value of those
octets is predetermined by the bench configuration. Only the first two
octets of the address mask matter for this (and most other) addressing
schemes.
We will add addressing schemes that do the right thing. Meanwhile, you
will be better off setting the server addresses manually as I
suggested earlier (using server_side.hosts).
> sure. I'm just starting. I'll get more dexterous as I go. I'm just
> trying to get the default configuration working right now to see
> if the L4 box I have fits your setup.
I hope you will find SrvLB workloads easier to use. The no-balancing
workload was released on-demand and is less polished. All workloads
require some effort if you start from scratch (and use Linux) though.
Good luck,
Alex.
This archive was generated by hypermail 2b29 : Tue Jul 10 2001 - 12:00:19 MDT