>
>
>Timm Murray (hardburn at runbox.com <mailto:hardburn at runbox.com>) wrote:
>
>>/ Those "multiple instances" are just threads.  Just the JVM alone has a fe=
>/w=20
>>/ threads of its own.  111 sounds a bit excessive, though.
>/
>Hell, no.  That's a LIGHT load.
>
>freenet at dwarf <mailto:freenet at dwarf>:~$ ps xw | grep java | wc
>    160    1119    9418
>
>(And that's a node that's only been up for a few minutes.)
>
>--=20
>Greg Wooledge                  |   "Truth belongs to everybody."
>greg at wooledge.org <mailto:greg at wooledge.org>              |    - The Red 
>Hot Chili Peppers
>http://wooledge.org/~greg/ <http://wooledge.org/%7Egreg/>     |
>
>  
>

You are both misinterpreting what I am saying. What I saw were 111 
threads for *each* connection.
The output of "lsof -i | grep java | wc -l " gave "6937". While a total 
of 111 threads wouldn't
have made me look twice, over 6 thousand kind of concerned me as to the 
state of a build that
is to be considered "stable" soon.

The only oddities about my system is that I am running with a minimal 
freenet.conf.

Only:

ipAddress=my.ip.address
listenPort=21872
seedNodes=seednodes.ref

to be precise.

I added FECTempDir and mainport.params.servlet.1.params.tempDir later
after my node complained about them not being there. I just would have
thought that there would have been some reasonable defaults that would
have prevented run-away thread counts or the limits are not being applied
correctly.

On a positive note, it was still fulfilling my requests eventhough it 
had a ton
of threads going. It was just not so perky about it.

Mike


_______________________________________________
devl mailing list
devl at freenetproject.org
http://hawk.freenetproject.org/cgi-bin/mailman/listinfo/devl

Reply via email to