Title: Re: [Oscar-devel] Problem with fc4, trunk and oda
I just tested this with "make install" from r5037 and I got the same problem as before - yum-oscar and yume will be uninstalled when I execute ./install_cluster and it can't continue.  I set OSCAR_VERBOSE=5 and provide a link out to the log:

http://www.bcgsc.ca/downloads/oscar/fc4-r5037-yum-error.txt
 
Basically I guess the problem is yum-oscar provides "yum", and when it tries to uninstall yum, it selects yum-oscar.  yume gets pulled in too because it requires yum.  This is an "exception" to the norm as all other distroes either has a new enough yum (eg. FC5 and SUSE10.0) or does not have yum at all (RHEL3, RHEL4).
 
Did something changed which may have caused this problem?  I believe this wasn't an issue on Fedora Core 3 when we had the old yum-oscar, and it had !yum in prereqs.cfg...
 
Anyways, I could still bring up the wizard if I re-run install_cluster, it sort of alternates between failing and succeeding... :-)
 
The end result is that trunk did install "successfully" on Fedora Core 4 (except for the above mentioned yum issues), and all tests passed.  Didn't encounter any ssh/MySQL issues.
 
I kickstarted the head and run "make install" - I deselect pfilter and everything else is quite standard...
 
Thomas, do you think you can test FC4 on your hardware and report back any issues?
 
Thanks,
 
Bernard


From: DongInn Kim [mailto:[EMAIL PROTECTED]
Sent: Thu 22/06/2006 18:21
To: [EMAIL PROTECTED]
Cc: Bernard Li; [email protected]
Subject: Re: [Oscar-devel] Problem with fc4, trunk and oda

Hi,

I have tried to test oscar r5035 on FC4-x86 and everything seems to work
fine for me.
Basically i tried three test. I know 1 and 2 are the same but I wanted
to make sure that "yum upate yum" works fine for oscar installation on
fc4 and the yum rpm is the same.
1. Download yum-2.4.1-1.fc4.noarch.rpm and update yum with this (rpm
-Uvh yum-2.4.1-1.fc4.noarch.rpm).
    And then copy it to $OSCAR_HOME/packages/yume/distro/fc4-i386
   (For those of you who use oscar nightly tarball, you should copy it
to /tftpboot/oscar/fc-4-i386/ too but if you install oscar from the svn
check out and do "make install" or "make test", then you don't need to
copy it)

2. "yum update yum" and then copy yum-2.4.1-1.fc4.noarch.rpm to
$OSCAR_HOME/packages/yume/distro/fc4-i386

3. Use yum-oscar-2.4.3-1.noarch.rpm that Geoffrey checked in today(6/22).
    I copied it to $OSCAR_HOME/packages/yume/distro/fc4-i386 and then I
modified yume/prereq.cfg by adding the following lines
    [fedora*:4:*]
    perl-IO-Tty
    python-urlgrabber
    python-elementtree
    !yum
    yum-oscar
    createrepo-0.4.3-5.1e
    yume

I have not run into the MySQL error and the weird ssh problems at all. :-)
The only weird thing is that the torque rpms seem to not work properly
at the testing. So I just tried to use the new torque rpms rebuilt on my
test machine and then the testing was fine too.
I will check in my new torque rpms and prereq.cfg shortly.

[EMAIL PROTECTED] yume]# cat /etc/redhat-release
Fedora Core release 4 (Stentz)
[EMAIL PROTECTED] yume]# uname -a
Linux oscar-fc4 2.6.11-1.1369_FC4 #1 Thu Jun 2 22:55:56 EDT 2005 i686
i686 i386 GNU/Linux

Regards,

- DongInn

Performing root tests...
Maui service check:maui                                        [PASSED]
TORQUE node check                                              [PASSED]
TORQUE service check:pbs_server                                [PASSED]
/home mounts                                                   [PASSED]

Preparing user tests...
Performing user tests...
SSH ping test                                                  [PASSED]
SSH server->node                                               [PASSED]
SSH node->server                                               [PASSED]
PVM (via TORQUE)                                               [PASSED]
Open MPI (via TORQUE)                                          [PASSED]
MPICH (via TORQUE)                                             [PASSED]
LAM/MPI (via TORQUE)                                           [PASSED]
Ganglia setup test                                             [PASSED]
Ganglia node count test                                        [PASSED]
TORQUE default queue definition                                [PASSED]
TORQUE Shell Test                                              [PASSED]

Run APItests...

Running Installation tests for pvm
[PASS]       2006-06-22 18:05:00   pvmd-path-ls.apt
[PASS]       2006-06-22 18:05:00   envvar-pvm_arch.apt
[PASS]       2006-06-22 18:05:00   envvar-pvm_root.apt
[PASS]       2006-06-22 18:05:00   envvar.apb
[PASS]       2006-06-22 18:05:00   pvmd-path-which.apt
[PASS]       2006-06-22 18:05:00   modulecmd-path-ls.apt
[PASS]       2006-06-22 18:05:00   pvm-module-list.apt
[PASS]       2006-06-22 18:05:01   pvm-module-show-pvm_rsh.apt
[PASS]       2006-06-22 18:05:01   pvm-module-show-pvm_arch.apt
[PASS]       2006-06-22 18:05:01   pvm-module-show-pvm_root.apt
[PASS]       2006-06-22 18:05:01   pvm-module-show.apb
[PASS]       2006-06-22 18:05:01   pvm-module.apb
[PASS]       2006-06-22 18:05:01   install_tests.apb

...Hit <ENTER> key to exit...

[EMAIL PROTECTED] wrote:
> Quoting Bernard Li <[EMAIL PROTECTED]>:
>
>  
>> Hi Geoffroy:
>>
>> Can you please provide how you reached this state:
>>
>> 1) How did you invoke the Wizard
>>    
>
> install_cluster since i said i had to re-launch install_cluster
>
>  
>> 2) Steps you executed to reach this state
>>    
>
> Just a normal installation process, the problem seems to happen after few hours
> (over the night in my case).
>
>  
>> Perhaps easiest way is to post your oscarinstall.log.
>>
>> I have deleted nodes before with FC5 and didn't have any issues, haven't
>> tried with FC4 though as far as I remember.
>>
>> P.S. Perhaps mysqld wasn't running when you ran "Delete Nodes"?
>>    
>
> I do not think it is a problem with the delete node step... MySQL freaked out.
> Unfortunately it seems i do not have mysql logs.
>
>  
>> Cheers,
>>
>> Bernard
>>
>>    
>>> -----Original Message-----
>>> From: [EMAIL PROTECTED]
>>> [mailto:[EMAIL PROTECTED]] On Behalf
>>> Of [EMAIL PROTECTED]
>>> Sent: Wednesday, June 21, 2006 13:09
>>> To: [email protected]
>>> Subject: [Oscar-devel] Problem with fc4, trunk and oda
>>>
>>> Impossible to cleanly delete my nodes, it seems that ODA
>>> quits periodically
>>> (actually it is more a MySQL issue). If i re-launch
>>> install_cluster everything
>>> seems to be back to normal.
>>> Did someone have this problem before?
>>>
>>> --> Update Wizard Env (as needed)
>>> DBD::mysql::st execute failed: MySQL server has gone away at
>>> /data/src/oscar/lib/OSCAR/oda.pm line 798.
>>> DBD::mysql::st execute failed: MySQL server has gone away at
>>> /data/src/oscar/lib/OSCAR/oda.pm line 798.
>>> DBD::mysql::st execute failed: MySQL server has gone away at
>>> /data/src/oscar/lib/OSCAR/oda.pm line 798.
>>> DBD::mysql::st execute failed: MySQL server has gone away at
>>> /data/src/oscar/lib/OSCAR/oda.pm line 798.
>>>      
>>>>> Turning off generic services
>>>>> Executing post_clients phase
>>>>>          
>>> No ipchains on this computer
>>> No modification to do for ipchains
>>> --> About to run
>>> /data/src/oscar/packages/sis/scripts/post_clients for sis
>>> using ODA to read the OSCAR database for node and adapters
>>> information ...
>>> reading SIS database for node and adapters information ...
>>> Node oscarnode1 is listed in the OSCAR database as using sis
>>> as the installer,
>>> but node oscarnode1 is NOT in the SIS database, DELETING node
>>> oscarnode1 from
>>> the OSCAR database ...
>>> Done deleting node oscarnode1 from the OSCAR database.
>>> --> About to run
>>> /data/src/oscar/packages/switcher/scripts/post_clients for
>>> switcher
>>> --> About to run
>>> /data/src/oscar/packages/c3/scripts/post_clients for c3
>>> --> About to run
>>> /data/src/oscar/packages/kernel_picker/scripts/post_clients for
>>> kernel_picker
>>>      
>>>>> Executing post_install phase
>>>>>          
>>> --> About to run
>>> /data/src/oscar/packages/ganglia/scripts/post_install for
>>> ganglia
>>> [ganglia] Ganglia gmond configuration file modified,
>>> re-starting daemon...
>>> Shutting down GANGLIA gmond:                               [  OK  ]
>>> Starting GANGLIA gmond:                                    [  OK  ]
>>> editing /etc/gmetad.conf
>>> match: gridname\s+.*
>>> match: data_source\s+.*
>>> [ganglia] Ganglia gmetad configuration file modified,
>>> re-starting daemon...
>>> Shutting down GANGLIA gmetad:                              [  OK  ]
>>> Starting GANGLIA gmetad:                                   [  OK  ]
>>> [ganglia] Starting up apache...
>>> Stopping httpd:                                            [  OK  ]
>>> Starting httpd:                                            [  OK  ]
>>> [ganglia] Ganglia page is located at
>>> http://ft_headnode.ornl.gov/ganglia/
>>> --> About to run
>>> /data/src/oscar/packages/loghost/scripts/post_install for
>>> loghost
>>> ************************* oscar_cluster *************************
>>> --> About to run
>>> /data/src/oscar/packages/switcher/scripts/post_install for
>>> switcher
>>> Setting default for tag mpi ("lam-7.1.2")
>>> Tag "mpi" does not seem to exist yet.  Skipping.
>>> --> About to run
>>> /data/src/oscar/packages/ntpconfig/scripts/post_install for
>>> ntpconfig
>>> Shutting down ntpd:                                        [  OK  ]
>>> Starting ntpd:                                             [  OK  ]
>>> ************************* oscar_cluster *************************
>>> --> About to run
>>> /data/src/oscar/packages/opium/scripts/post_install for opium
>>> Could not find template for file switcher.ini, distro redhat
>>> If this contains distro-specific lines, please create a template
>>> in directory /opt/sync_files/templates
>>> image:
>>> $VAR1 = 'oscarimage';
>>> ---------------
>>> cpush returned 65280 on subcluster oscarimage
>>> Could not find template for file gshadow, distro redhat
>>> If this contains distro-specific lines, please create a template
>>> in directory /opt/sync_files/templates
>>> image:
>>> $VAR1 = 'oscarimage';
>>> ---------------
>>> cpush returned 65280 on subcluster oscarimage
>>> image:
>>> $VAR1 = 'oscarimage';
>>> ---------------
>>> cpush returned 65280 on subcluster oscarimage
>>> image:
>>> $VAR1 = 'oscarimage';
>>> ---------------
>>> cpush returned 65280 on subcluster oscarimage
>>> image:
>>> $VAR1 = 'oscarimage';
>>> ---------------
>>> cpush returned 65280 on subcluster oscarimage
>>> --> About to run
>>> /data/src/oscar/packages/oda/scripts/post_install for oda
>>> generating the /etc/odaserver file on all oscar clients
>>> . /etc/profile.d/c3.sh && cexec 'echo oscar_server > /etc/odaserver'
>>> ************************* oscar_cluster *************************
>>> Cluster setup complete!
>>>      
>>>>> Re-starting generic services
>>>>> Re-starting server services
>>>>> Updating C3 configuration file
>>>>> Re-starting client services on remaining nodes
>>>>>          
>>> Use of uninitialized value in concatenation (.) or string at
>>> /data/src/oscar/lib/OSCAR/Network.pm line 43.
>>> Use of uninitialized value in concatenation (.) or string at
>>> /data/src/oscar/lib/OSCAR/DelNode.pm line 197.
>>> Use of uninitialized value in string eq at
>>> /data/src/oscar/lib/OSCAR/DelNode.pm
>>> line 199.
>>> Running mkdhcpconf
>>> Option interface requires an argument
>>> usage: mkdhcpconf <options>
>>>
>>>   options
>>>     --outfile <name>        the filename to write the output to
>>>     --interface <name>      the server's network interface to
>>> listen on.
>>> (default, eth0)
>>>     --bootfile <file>       the bootfile to serve to clients.
>>> (default,
>>> /tftpboot/pxelinux.bin)
>>>     --gateway <host>        the default route for the machines
>>>     --domain <domain>       the domain of the machines
>>> (default, server domain)
>>>     --multicast <yes>       to enable multicastng
>>>     --nameservers <servers> a comma delimited list of nameservers.
>>>     -v, --verbose           massive verbose output
>>>
>>>
>>> Tk::Error: Failed to run mkdhcpconf -o /etc/dhcpd.conf --interface=
>>> --gateway=192.168.1.1 at
>>> /usr/lib/perl5/vendor_perl/5.8.6/i386-linux-thread-multi/Tk.pm
>>>  line 247
>>>  Carp::croak at /usr/lib/perl5/5.8.6/Carp.pm line 269
>>>  OSCAR::DelNode::delnodes at
>>> /data/src/oscar/lib/OSCAR/DelNode.pm line 204
>>>  Tk callback for .toplevel.button
>>>  Tk::__ANON__ at
>>> /usr/lib/perl5/vendor_perl/5.8.6/i386-linux-thread-multi/Tk.pm
>>> line 247
>>>  Tk::Button::butUp at
>>> /usr/lib/perl5/vendor_perl/5.8.6/i386-linux-thread-multi/Tk/Bu
>>> tton.pm line 111
>>>  <ButtonRelease-1>
>>>  (command bound to event)
>>>
>>>
>>> All the advantages of Linux Managed Hosting--Without the Cost
>>> and Risk!
>>> Fully trained technicians. The highest number of Red Hat
>>> certifications in
>>> the hosting industry. Fanatical Support. Click to learn more
>>> http://sel.as-us.falkag.net/sel?cmd=lnk&kid=107521&bid=248729&
>>> dat=121642
>>> _______________________________________________
>>> Oscar-devel mailing list
>>> [email protected]
>>> https://lists.sourceforge.net/lists/listinfo/oscar-devel
>>>
>>>      
>
>
>
> Using Tomcat but need to do more? Need to support web services, security?
> Get stuff done quickly with pre-integrated technology to make your job easier
> Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo
> http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642
> _______________________________________________
> Oscar-devel mailing list
> [email protected]
> https://lists.sourceforge.net/lists/listinfo/oscar-devel
>  

Using Tomcat but need to do more? Need to support web services, security?
Get stuff done quickly with pre-integrated technology to make your job easier
Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo
http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642
_______________________________________________
Oscar-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/oscar-devel

Reply via email to