From: Hironori Ito (hito@rcf.rhic.bnl.gov)
Date: Wed Jun 18 2003 - 15:04:32 EDT
Ok. After the rcf upgrade meeting, I would say that there is a good chance that all those software will be installed as you hoped. If you have a software you really want to install, please let me know as soon as possible. The size of root pertition is 5GB (probably). Therefore, you should be able to ask alot of things (of course, only things you and everybody else need). The time frame for this upgrade is about the mid to the end of August. I highly suggest to test many of softwares in RH8 and RH9 play machines. One software I did not test is cernlib. Someone should test cernlib. Hiro Hironori Ito wrote: > > ------------------------------------------------------------------------ > > Subject: > Re: OS upgrade xv,xfig > From: > Hironori Ito <hito@rcf.rhic.bnl.gov> > Date: > Wed, 18 Jun 2003 14:00:10 -0400 > To: > Bj|rn Hallvard Samset <b.h.samset@fys.uio.no> > > > Hello. About xv and xfig as well as latex, (although I do not > underestand why someone with right mind (and not enormous patience) want > to use it rcas over looooooong network,) I may ask it to rcf. But, as > Flemming said it, rcf probably will not install it. (It is available at > rcf2.) > > About swish-e, I also heard about it somewhere. I might try it to our > list server. (well to our files in list server) > > Hiro > > Bj|rn Hallvard Samset wrote: > >> On Wed, 18 Jun 2003, Flemming Videbaek wrote: >> >> Hi Flemming et Al. >> >> >>> I would mention that I also find it inconvenient that xv, and the fig >>> generationis not on the rcas - on the other hand the RCF policy is to >>> have such program available on rcf2 where they de facto are. I do >>> specificaly recall a discussion way back, they would only maintain latex >>> on rcf2, but this was a long time ago and may not even be correct. >> >> >> >> As I've said before, I strongly support latex, xv and xfig on rcas, >> mainly >> for the reasons mentioned by Christian but also because it makes >> interactive work at BNL much easier (less moving of small files). >> >> >>> On the list server. I do not think we should have the private list index >>> by google if this means they must be public ! That why it is on the >>> private page. >> >> >> >> I've recently looked into a nice little package called swish-e [1] which >> makes indexing e.g. our list servers very easy. I have it working on the >> command line (default mode), but so far I've been unable to get a cgi >> script or the likes to work (I don't have much experiece with this, so >> I'm >> learning as I go). >> >> An example of a swish-e search: >> >> rcas0028(8) ~/swish/indices>swish-e -f brahms-lists.index -w dAu -m 10 >> # SWISH format: 2.4.0-pr1 >> # Search words: dAu >> # Removed stopwords: >> # Number of hits: 57 >> # Search time: 0.001 seconds >> # Run time: 0.017 seconds >> 1000 ../test/brahms-dev-l/2607.html "brahms-dev-l: dAu setup and analysis >> question" 4829 >> 979 ../test/brahms-dev-l/2611.html "brahms-dev-l: Re: dAu setup and >> analysis question" 5522 >> 979 ../test/brahms-dev-l/2613.html "brahms-dev-l: Re: dAu setup and >> analysis question" 6072 >> 955 ../test/brahms-dev-l/2608.html "brahms-dev-l: Re: dAu setup and >> analysis question" 4500 >> 944 ../test/brahms-dev-l/2615.html "brahms-dev-l: Re: dAu setup and >> analysis question" 6301 >> 908 ../test/brahms-dev-l/2525.html "brahms-dev-l: bb vertex resolution in >> dAu run" 3735 >> 894 ../test/brahms-l/2436.html "brahms-l: Fw: Abstract of Phobos paper on >> dAu results" 4918 >> 879 ../test/brahms-l/2260.html "brahms-l: remaining dAu Run Plan" 3858 >> 879 ../test/brahms-dev-l/2521.html "brahms-dev-l: zdc vertex >> resolution in >> dAu run" 3919 >> 864 ../test/brahms-l/2439.html "brahms-l: update on dAu analysis" 3465 >> >> I can also tag it to search in file header/body/whatever: >> -t : tags to search in - specify as a string >> "HBthec" - in Head|Body|title|header|emphasized|comments >> >> In an html interface this will all of course be clickable etc. Have a >> look >> here: >> http://www.sdcc.bnl.gov/brahms/private/test/ >> for a non-working example of what it may look like. (Source from [2]) >> Or a >> working version on another site here: >> http://www.webaugur.com/search/ >> >> Updating the index can be set up as a cron job run by e.g. brahmlib or >> bramreco, so once the thing is up it should be self-maintaining. I can't >> guarantee that this will function the way we want, but to me it seems >> like >> an easy way of getting at least close. I'll keep looking at it, unless >> someone has a better idea. >> >> Ping :-) >> >> -Bjorn >> >> [1] http://swish-e.org/ >> [2] http://www.webaugur.com/wares/sws.html >> >> >>> ----- Original Message ----- >>> From: "Christian Holm Christensen" <cholm@hehi03.nbi.dk> >>> To: <brahms-dev-l@bnl.gov>; <hito@rcf.rhic.bnl.gov> >>> Sent: Wednesday, June 18, 2003 4:44 AM >>> Subject: Re: OS upgrade >>> >>> >>> | Hi Hiro, >>> | >>> | Hironori Ito <hito@rcf.rhic.bnl.gov> wrote concerning >>> | Re: OS upgrade [Tue, 17 Jun 2003 18:16:19 -0400] >>> | ---------------------------------------------------------------------- >>> | > Hello. About rcas software. >>> | > >>> | > Who use xv, xfig and latex over the long network? I wouldn't do >>> it. Do >>> | > we realy ask for this? (Note: rcf only wants to install the same >>> image >>> | > to all rcas including othere rhic experiments.) >>> | >>> | LaTeX is a good idea, as we use that for generating DVI and PS >>> | versions of `The Hitchhikers Guide to BRAT'. >>> | >>> | LaTeX2Html is a good idea too, as that's used for the HTML version of >>> | the above guide. >>> | >>> | Transfig would be nice too, as you may need that for some figures in >>> | the above mentioned Guide. XFig it self isn't really needed unless >>> | local users want to make drawings (like class diagrams to document >>> | code :-) >>> | >>> | I see no reason why you wouldn't use LaTeX over a network connection. >>> | After all, TeX does not need an X connection at all. >>> | >>> | I'd also recommend Doxygen. It's very cool for documenting code. In >>> | fact, I'd recommend that we switched to Doxygen for the automaticaly >>> | generated HTML documentation of BRAT, as it is much more capable than >>> | THtml. >>> | >>> | Doxymacs is cool too. >>> | >>> | > Flemming Videbaek wrote: >>> | > > I compiled the messages from Bjoern+Christian into this kind of >>> list. >>> | > > >>> | > > Since I will probably not make it to (all) of the meeting I >>> hope/assume >>> | > > Hiro can go. >>> | > > >>> | > > >>> | > > >>> | > > Other comments- >>> | > > >>> | > > Kris have you had a chance to checkout the RH9.0 machine-oops I >>> guess >>> | > > not since the >>> | > > >>> | > > message about the RH9 was probably not ever forwarded to you. >>> | > > >>> | > > >>> | > > >>> | > > | >>> | > > | The respective Liaisons are encouraged to test (or ask >>> | > > | your users to test) these 3 different RH Linux flavors >>> | > > | (rplay1 -> RH 7.3, rplay15 -> RH 8 and rplay10 -> RH 9) >>> | > > | in the next few weeks (the earlier the better). At the >>> | > > | meeting, each experiment will be asked for their OS >>> | > > | upgrade wishlist (flavor, extra software packages, etc). >>> | > > | Since USATLAS uses the LSF shared queues in the RHIC Linux >>> | > > | Farm, is is encouraged to attend in order to understand >>> | > > | how the RHIC OS upgrade will impact its access to the >>> | > > | RHIC Linux Farm. >>> | > > | >>> | > > | I will make a short presentation at the meeting to >>> | > > | summarize our upgrade plan and schedule. Discussion of >>> | > > | the experiments' needs will follow. >>> | > > >>> | > > Flemming >>> | > > >>> | > > >>> | > > >>> | > > Bjorn (sys op at NBI) and my general experience is, that one >>> should >>> | > > wait for Red Hat x.1 rather than jumping into Red Hat x.0 straight >>> | > > away, as x.0 always has child-diseases. That said, we should >>> get onto >>> | > > a newer system ASAP. >>> | > > >>> | > > >>> | > > >>> | > > >> As I hear RH 9.0 is in reality RH 8.1 thus should be ok >>> | >>> | What! I don't think that's true. Red Hat 9.0 is a thing for Private >>> | users, so the money people at Red Hat can push Advanced Server. A >>> | pure money scam - another good reason to switch to Debian or even >>> | Gentoo. >>> | >>> | >>> | > > Compilers: >>> | > > * GCC 3.2 or 3.3 (probably available at that time) >>> | > > * Intel C++ compiler >>> | > > * GDB >>> | > > >>> | > > >>> | > > >>> | > > scripting >>> | > > >>> | > > * Perl an obvious must >>> | > > * python2.2, >>> | > > >>> | > > >>> | > > >>> | > > Libraries: >>> | > > * XFree86 4.2.x run-time and developer libraries + headers >>> | > > (in particular the OpenGL and libfreetype stuff) >>> | > > * libmysqlclient (what ever version) >>> | > > * LDAP for GRID solutions >>> | > > * GSL (Gnu Scientific library) >>> | > > >>> | > > >>> | > > >>> | > > Tools >>> | > > >>> | > > * The MS True Type Fonts installed somewhere on the system. >>> | > > * Autotools: >>> | > > - Autoconf 2.53+ >>> | > > - Automake 1.7+ >>> | > > - Libtool 1.5+ >>> | > > >>> | > > >>> | > > >>> | > > could we also add utility programs like >>> | > > >>> | > > >> (request for making overseas work easier, except of course >>> these >>> | > > services are available on rcf2 ). >>> | > > >>> | > > >>> | > > * latex >>> | > > * xv >>> | > > * xfig >>> | > > >>> | > > >>> | > > >>> | > > >>> | > > >>> | > > Services: >>> | > > * MySQL server on brahms-db0 >>> | > > * rootd - that's a can of worms to say the least. To do that >>> really >>> | > > effectively, we should get the RCF guys to have `rootd' >>> controlled >>> | > > by `xinetd' (see the `rootd' man page, and the RPM stuff in >>> the >>> | > > ROOT source tree). >>> | > > >>> | > > >> This is how it is done, apart from the fact it points not >>> to a /uar/.. >>> | > > >>> | > > >> but a brahmlib version. >>> | >>> | If it doesn't point to /usr, it should point to the AFS version. >>> | >>> | > > * LSF or OpenPBS. (will likely have condor) >>> | >>> | Condor is sitting on top of LSF or OpenPBS. >>> | >>> | > > >>> | > > >> >>> | > > * LDAP >>> | > > * CVS server. Rather than insisting on people using AFS to >>> access >>> | > > our CVS repositories, we could use a CVS server. >>> | > > >>> | > > >> -- Certainly not on all machines? >>> | >>> | No, on some designaed machine, say brahms-cvs (take one of the older >>> | ones). >>> | >>> | > > >>> | > > Other stuff: >>> | > > * The MS True Type Fonts installed somewhere on the system. >>> | > > * Autotools: >>> | > > - Autoconf 2.53+ >>> | > > - Automake 1.7+ >>> | > > - Libtool 1.5+ >>> | > > >>> | > > >>> | > > >>> | > > >>> | > > >>> | > > The web search engine is as I realized since Betty left our own >>> problem >>> | > > not RCF since the >>> | > > >>> | > > messages end up under our control on the br_list user account. >>> | >>> | Then install HtDig or Glimpse on the web-server, or ask Google to >>> | index the pages. They offer that service freely to academia. >>> | >>> | Yours, >>> | >>> | ___ | Christian Holm Christensen >>> | |_| | ------------------------------------------------------------- >>> | | | Address: Sankt Hansgade 23, 1. th. Phone: (+45) 35 35 96 91 >>> | _| DK-2200 Copenhagen N Cell: (+45) 24 61 85 91 >>> | _| Denmark Office: (+45) 353 25 305 >>> | ____| Email: cholm@nbi.dk Web: www.nbi.dk/~cholm >>> | | | >>> | >>> >>> >> >> >> -- >> Bjørn H. Samset Tlf: 22856465/92051998 >> Dr. gradsstudent, tungionefysikk Adr: Schouterrassen 6 >> Fysisk institutt, Universitetet i Oslo 0573 Oslo >> \|/ >> ----------------------------> -*- <----------------------------- >> /|\ > > >
This archive was generated by hypermail 2.1.5 : Wed Jun 18 2003 - 15:07:08 EDT