Re: (FWD) Results of stress testing the current version of the Onion Router

[Bruce Montrose <montrose@itd.nrl.navy.mil>]

Thanks for useful info.

By down I mean that the httpap was not talking to anybody, so i suppose the 
problem may be confined to that process and the actual "or" processes may 
have weathered the storm.

At 01:35 AM 8/29/2002 -0400, you wrote:
> > ----- Forwarded message from Bruce Montrose <montrose@itd.nrl.navy.mil> 
> -----
> > When the test began data seemed to be flowing well through the onion
> > routers, but as the load increased things deteriorated quickly. After the
> > test completed, I attempted to bring up my browser which I setup to proxy
> > through the onion router and it failed. The onion_routers are still
> > running, but they appear to be hung.
> >
> > I will attempt to duplicate this problem with LogLevel = debug to see if I
> > can get any clues as to why the onion routers are hanging. FYI, I ran this
> > test with LogLevel = err and I saw no output in the log files for the 
> onion
> > routers. I don't know if this is because there were no errors detected by
> > the onion routers or if the output buffers were never flushed to the 
> log files.
>
>Running the ORs with -l debug will make them *much* less efficient. But
>maybe that will just force the problem to show up earlier. :)
>
>Also try using strace -- figure out the pid of the process, and "strace
>-p pid". That should give you a sense of what system calls the program
>is up to (perhaps it's blocking on something, or perhaps it's in a loop).
>
>You can also attach to the process with gdb, via "gdb or pid". That
>should let you look at things in more detail.
>
>What exactly do you mean by 'failed'? Did it succeed at connecting to
>the OR but got to response? Or was the connection itself down.
>
>--Roger