gomd-devel
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [gomd-devel] <DAEMON> gomd testing


From: Matthias Rechenburg
Subject: Re: [gomd-devel] <DAEMON> gomd testing
Date: Sat, 26 Jul 2003 17:34:01 +0200
User-agent: KMail/1.4.3

Ciao my friend,

On Samstag 26 Juli 2003 14:16, Gian Paolo Ghilardi wrote:
> Hi Matt.
> How're you?

i am fine, thx, what about you ?
.. just cleaned-up our basement with my wife...
..... and found my old and self-made "cyber-gloves" :))
about 100 years ago i built some gloves connected to an
old keybord and you could use it for e.g. Doom and Decent
to fly + run around by typing with your fingers on a desk.
.... will post a pic soon, hehehe.
...... maybe i can/will use it for my cluster-nodes management now ....lol

>
> > Ciao JP,
> >
> > here the short summary of my testing of the latest gomd+libgomd sources :
> >   :))  very nice !
> >
> > now a bit more detailed :
> > - had a bit trouble to get the libgtop-2.0.2 (latest) working on my
> > Redhat system. The libgtop-1.x.x was installed but there is no version
> > 2.x available as an rpm .... so i compiled myself and .....
> >   ..... the "make" resulted in an endless loop and did not start
> >   compiling :/  ....... ok, found out why : the autoconf add 2 spaces
> >   in each Makefile where no spaces are allowd (just tabs). I had to
>
> correct
>
> >   every Makefile in every subdir and then it compiled ok.
> >   The questions are :
> >   Do we really need libgtop-2.x.x or can we reduce the deps
> >   to libgtop-[any-version] ?
>
> Uhm... Libgtop2 contains a lot of new features.
> Maybe I'm wrong but the 2.0 version of libgtop was entirely rewritten.
> However I'll check. ;)
> Please notice libgtop2 is optional.

yep, but it is hard to use for the gomd-users (to my mind)
... maybe we can provide a "compile-able" libgtop-2 in 
our cvs  ... !?

>
> > - i do not really like the locate + find in the compile.sh ;)
> >   It simply takes too long and it not safe (to my mind).
>
> Uhm...Unsafe? Why?

... it may find 2 versions, which it will use ?
anyways to my mind telling the user that he needs
commoncpp+libgtop is enough. If he/she does not
install this first he/she will see that the compilation
failed because of the missing component.

>
> >   I suggest just to remove those library-checks and
> >   let the compile.sh only doing the compile. We can e.g.
> >   check the exit status of the make and tell the user that
> >   some libs are missing.
>
> It's an idea but... what if compile.sh asked the user something like (in
> compile.sh):
> "Do you want to include extra informations support (requires libgtop2
> installed): [y/N]"?

;) yep, you are right. ..... to my mind we should use the libgtop2 
anyways ;) those infos from it are very usefull for client apps.

>
> > That's all ;)
> > Everything else worked like a charme in my testing !
> > I like the test1.sh+test2.sh and also the client-example.app.
> > The new values in the gomd (from libgtop) are VERY usefull !
> > ... and thanks to JP's function-numbering we are not
> > far away from having all this new informations available in the
> > libgomd too i guess.
>
> Matt. About the function ids, I'd like to (re)define some ranges.
> For examples:
> - functions with id < 100 require special privileges
> - functions with ranges between 101 and 200 can be requested only on the
> local node (or with special privileges)
> - function with id >201 don't require special privileges
> Comments?

go ahead, no problem with that.

> Matt: can you  (when you can/want, obviously) arrange the function ids?
> Thanks.
> They are (logically) untidy. ;)
>
> > We should now put the hardcoded ip-adresses in omInfo.cpp in the
> > gomd.conf and the ip-addresses in the libgomd-code into a "libgomd.conf"
> > .. ... or maybe in the gomd.conf too. Let's discuss ;)
>
> I'm working to avoid hard coded stuff. ;)

:)

>
> > many thanks JP, really good work.
> > Also i like our project page + diagrams :)   ... and the README...
> > gomd->fun();
> >
> > have a nice weekend,
>
> Also to you, Matt.
> Thanks a lot for your testing... ;)

welcome, the gomd is fun. 
i promise that i will write a small qt-example app as soon
as some more functions are working in the libgomd.

>
>
> CU.
> Bye.
>
> <rejected>
>
>
>
>
> _______________________________________________
> gomd-devel mailing list
> address@hidden
> http://mail.nongnu.org/mailman/listinfo/gomd-devel

all the best,

Matt
-- 
E-mail  :  address@hidden
www     : http://www.openmosixview.com
an openMosix-cluster management GUI

If you're not part of the solution, you're part of the precipitate.





reply via email to

[Prev in Thread] Current Thread [Next in Thread]