On Thu 21 October 2004 19.43, Bernard Li wrote:
> [moved to oscar-devel list]

Yes, you're right, so here I am :-)

So maybe I should start with a short self-intro:
I work as UNIX admin at Karolinska Instituttet in Stockholm, split between a 
department called the Center for Genetics and Bioinformatics, and a service 
called the Bioinformatics and Expression Analysis Core Facility.

Other than day-to-day maintenance, I've got a pet project BioRPMS 
<http://apt.bea.ki.se> - this actually grew from an internal need for rpm's 
to keep our own machines in sync with each other. I do NOT claim to be an 
expert at rpmbuilding, but have done about 90 so far - they may be silly and 
primitive, though; the earlier ones make even me shake my head in disgusted 
amusement when I have to go back and update them. 

My cluster experience is somewhat limited: Upon taking the job, I inherited an 
old (1998) DYI beowulf, and couldn't make heads or tails of it, so I ended up 
turning half of it into an oscar, which seemed to please the users. 

Our next deployment involves two small clusters, each with a dual opteron 
frontend and a few single opteron nodes. One cluster for powering our web 
services, and one for developing them. I had nearly given up on oscar when I 
saw Bernard's mail last week about oscar-4.0, and so decided to see if I 
could in some small way help to make it happen. I'd guess that specfiles and 
rpms is where I could be of must use, if any.

> Are you encountering any problems installing the code from SVN under Fedora
> Core 2?  I have no problem using the RPMs provided in the
> /oscar/packages/perl-QT/RPMS directory (I installed on FC2 i386...).

After installing PerlQt from source, I ran configure again, and then 
discovered the perlqt pakages hiding inside....

I've tried to do a rough'n'dirty --rebuild on all the srpms on FC2 i386 - I 
figured that ought to work before addressing x86_64 specific issues. That was 
a bit disappointing, though.

[EMAIL PROTECTED] biobuild]$ more results.txt
hdf5: ./hdf5/SRPMS/hdf5-oscar-module-1.6.0-1.src.rpm - success
hdf5: ./hdf5/SRPMS/hdf5-oscar-1.6.0-1.src.rpm - failure
loghost: ./loghost/SRPMS/loghost-1.0-1.src.rpm - success
torque: ./torque/SRPMS/torque-1.0.1p5-2.src.rpm - failure
maui: ./maui/SRPMS/maui-oscar-3.2.5p2-7.src.rpm - failure
documentation: ./documentation/SRPMS/oscar-user-docs-html-2.0b1-1.src.rpm - 
success
documentation: ./documentation/SRPMS/oscar-install-docs-html-2.0b1-1.src.rpm - 
success
lam: ./lam/SRPMS/lam-oscar-7.0.6-1.src.rpm - success
lam: ./lam/SRPMS/lam-switcher-modulefile-7.0.6-1.src.rpm - success
lam: ./lam/SRPMS/libaio-oscar-0.3.93-5.src.rpm - success
switcher: ./switcher/SRPMS/modules-oscar-3.1.6-5.src.rpm - failure
switcher: ./switcher/SRPMS/env-switcher-1.0.13-1.src.rpm - success
perl-Qt: ./perl-Qt/SRPMS/perl-Qt-5.6-3.006-2.src.rpm - failure
perl-Qt: ./perl-Qt/SRPMS/perl-Qt-mdk-3.006-5.src.rpm - failure
perl-Qt: ./perl-Qt/SRPMS/perl-Qt-3.006-5.src.rpm - failure
perl-Qt: ./perl-Qt/SRPMS/perl-Qt-qt3-compat-1.0-2.src.rpm - success
sis: ./sis/SRPMS/systeminstaller-1.04-1.src.rpm - failure
sis: ./sis/SRPMS/systemimager-3.2.0-1.src.rpm - failure
sis: ./sis/SRPMS/systemconfigurator-2.0.9-1.src.rpm - failure
sis: ./sis/SRPMS/flamethrower-0.1.6-1.src.rpm - failure
kernel_picker: ./kernel_picker/SRPMS/kernel_picker-1.4-2.src.rpm - failure
c3: ./c3/SRPMS/c3-4.0.1-2.src.rpm - success
c3: ./c3/SRPMS/python2-2.1.1-1tjn.src.rpm - failure
oda: ./oda/SRPMS/oda-1.30-5.src.rpm - failure
mpich: ./mpich/SRPMS/mpich-ch_p4-gcc-oscar-1.2.5.10-2.src.rpm - success
ganglia: ./ganglia/SRPMS/ganglia-oscar-monitor-core-2.5.6-1.src.rpm - failure
sync_files: ./sync_files/SRPMS/sync_files-1.6-5.src.rpm - success
pfilter: ./pfilter/SRPMS/pfilter-1.705-1.src.rpm - failure
pfilter: ./pfilter/SRPMS/pfilter-1.699-1.src.rpm - failure
pvm: ./pvm/SRPMS/pvm-3.4.4+11-3.src.rpm - success
pbs: ./pbs/SRPMS/openpbs-oscar-2.3.16-11.src.rpm - failure
opium: ./opium/SRPMS/ssh-oscar-1.1-5.src.rpm - failure

Glancing briefly at the buildlogs, it looks as if some of the failures were 
caused by packages expecting to be built as root. As this contradicts what 
everybody else is telling me about rpm-building, I have to ask if this is a 
deliberate design choice? And if yes, then why?

Or maybe there's some clever chrooted environment setup that I need to build 
inside? I've tried working with mach, but there were too many unresolved 
issues, so I'm waiting/hoping for a new release.

I still feel that I need to get a grip on building on i386 first, then tackle 
x86_64.

> P.S. You need to be subscribed to oscar-devel list to post back, incase you
> aren't already...

Am now :-)

>
> > -----Original Message-----
> > From: [EMAIL PROTECTED]
> > [mailto:[EMAIL PROTECTED] On Behalf Of
> > Bent Terp
> > Sent: Thursday, October 21, 2004 2:02
> > To: [EMAIL PROTECTED]
> > Subject: [Oscar-users] branch-4.0 and FC2 x86_64
> >
> > Hi!
> >
> > I'd like to help with getting oscar 4.0 up and running on FC2
> > x86_64, and FC3 real-soon ;-)
> >
> > Is there a reason I can't find FC2 rpm's for PerlQt? Beyond
> > the usual "well, because nobody's built them".
> >
> > --
> > MVH / With kind regards,
> > � � � � Bent Terp <[EMAIL PROTECTED]>
> > � � � � System Administrator
> > � � � � Bioinformatics and Expression Analysis Core Facility
> > � � � � Karolinska Institute, Department of Biosciences
> > � � � � H�lsovagen 7-9, 141 57 Huddinge, Sweden
> >
> >
> > -------------------------------------------------------
> > This SF.net email is sponsored by: IT Product Guide on
> > ITManagersJournal
> > Use IT products in your business? Tell us what you think of
> > them. Give us
> > Your Opinions, Get Free ThinkGeek Gift Certificates! Click to
> > find out more
> > http://productguide.itmanagersjournal.com/guidepromo.tmpl
> > _______________________________________________
> > Oscar-users mailing list
> > [EMAIL PROTECTED]
> > https://lists.sourceforge.net/lists/listinfo/oscar-users

-- 
MVH / With kind regards,
� � � � Bent Terp <[EMAIL PROTECTED]>
� � � � System Administrator
� � � � Bioinformatics and Expression Analysis Core Facility
� � � � Karolinska Institute, Department of Biosciences
� � � � H�lsovagen 7-9, 141 57 Huddinge, Sweden


-------------------------------------------------------
This SF.net email is sponsored by: IT Product Guide on ITManagersJournal
Use IT products in your business? Tell us what you think of them. Give us
Your Opinions, Get Free ThinkGeek Gift Certificates! Click to find out more
http://productguide.itmanagersjournal.com/guidepromo.tmpl
_______________________________________________
Oscar-devel mailing list
[EMAIL PROTECTED]
https://lists.sourceforge.net/lists/listinfo/oscar-devel

Reply via email to