YAIM 4.0.0 Certification (second phase)

Developers testing

There is no testing done by the developers at this second stage.

Fresh install testing

rpm list

The list of rpms can be found in etics in /afs/cern.ch/project/gd/www/yaim:

• glite-yaim-core-4.0.0-9.noarch.rpm
• glite-yaim-clients-4.0.0-2.noarch.rpm
• glite-yaim-dcache-4.0.0-2.noarch.rpm
• glite-yaim-dpm-4.0.0-2.noarch.rpm
• glite-yaim-fts-4.0.0-5.noarch.rpm
• glite-yaim-lfc-4.0.0-3.noarch.rpm
• glite-yaim-myproxy-4.0.0-2.noarch.rpm

The main difference with the rpms from the first phase of certification is:

• config_gip and config_bdii in yaim core have been changed.
• node-info.def has changed for top level BDII and site BDII in yaim core.
• New fix in yaim core to the fix provided by Rolanda's to be able to finish the yaim script succesfully.
• Fixes in yaim core and yaim dpm for patch #1283 included also in this release.
• Fix in yaim core for APEL configuration in pbs.
• There are new Makefile and spec files in all the modules.

For a list of the bugs that this version of YAIM intends to fix, please see the following Savannah patches:

• patch #1238, YAIM 4.0.0 for the 3.0 repository
• patch #1239, YAIM 4.0.0 for the 3.1 repository

CERN testing report

General remark:

• For all resources, we need to uninstall lcg-info-generic and install glite-info-generic-2.0.2-1, besides, BDII needs to be upgarded to 3.9.1-3.
• When configuring BDII on all resources, it gives:

   INFO: Executing function: config_bdii 
/opt/glite/yaim/bin/../libexec/configure_node: line 100: [: : integer expression expected

3.0 UI

• When calling config_globus, it gives some error info:

/opt/globus/sbin/globus-initialization.sh: line 61: ./setup-globus-mds-gris:
No such file or directory
/opt/globus/sbin/globus-initialization.sh: line 62: ./setup-globus-gatekeeper:
No such file or directory
/opt/globus/sbin/globus-initialization.sh: line 63:
./setup-globus-gram-job-manager: No such file or directory
/opt/globus/sbin/globus-initialization.sh: line 64:
./setup-globus-gram-reporter-fork: No such file or directory
/opt/globus/sbin/globus-initialization.sh: line 65:
./setup-globus-gram-reporter-pbs: No such file or directory
/opt/globus/sbin/globus-initialization.sh: line 66:
./setup-globus-gram-reporter-condor: No such file or directory
/opt/globus/sbin/globus-initialization.sh: line 67:
./setup-globus-gram-reporter-lsf: No such file or directory
/opt/globus/sbin/globus-initialization.sh: line 68:
./setup-globus-job-manager-fork: No such file or directory
/opt/globus/sbin/globus-initialization.sh: line 69:
./setup-globus-job-manager-pbs: No such file or directory
/opt/globus/sbin/globus-initialization.sh: line 70:
./setup-globus-job-manager-condor: No such file or directory
/opt/globus/sbin/globus-initialization.sh: line 71:
./setup-globus-job-manager-lsf: No such file or directory

it's harmless, but annoying.

• when calling config_glite_ui_30,

setup-ssl-utils: Complete

..Done
WARNING: The following packages were not set up correctly:
        globus_trusted_ca_42864e48_setup-noflavor-pgm
Check the package documentation or run postinstall -verbose to see what
happened

this can be ignored although I list it here

3.0 WN

• when calling config_globus,

/opt/globus/sbin/globus-initialization.sh: line 65: ./setup-globus-gram-reporter-pbs: No such file or directory
/opt/globus/sbin/globus-initialization.sh: line 66: ./setup-globus-gram-reporter-condor: No such file or directory
/opt/globus/sbin/globus-initialization.sh: line 67: ./setup-globus-gram-reporter-lsf: No such file or directory
loading cache ./config.cache
checking for mpirun... /usr/bin/mpirun
updating cache ./config.cache
creating ./config.status
creating fork.pm
/opt/globus/sbin/globus-initialization.sh: line 69: ./setup-globus-job-manager-pbs: No such file or directory
/opt/globus/sbin/globus-initialization.sh: line 70: ./setup-globus-job-manager-condor: No such file or directory
/opt/globus/sbin/globus-initialization.sh: line 71: ./setup-globus-job-manager-lsf: No such file or directory

but it can be ignored.

3.0 LCG CE

• when calling config_fmon_client, it complaims:

   INFO: Executing function: config_fmon_client 
JM.conf file parsing: [FAILED]

Batch system log dir not defined in JM.conf!
JM.conf file parsing: [FAILED]

Batch system log dir not defined in JM.conf!

• when combining lcg CE and site_BDII on the same node, config_bdii creates schemas configuration file, /opt/bdii/etc/schemas, to include /opt/lcg/schema/ldap/SiteInfo.schema, however config_bdii_only used by LCG CE changed it to /opt/lcg/schema/openldap-2.0/SiteInfo.schema ---- Solution(?): go back to old MDS, for this config_bdii_only should be removed from CE function list and change config_globus to put back "GLOBUS_MDS=yes" in line 9

WMS+LB

The test is based on the current checkpoint release, 1251.

• node-info should be updated to use config_gip not config_gip_services, config_gip should be updated for WMS port and add LB node type.

LFC

• it calls config_bdii_only to configure BDII on LFC, but LFC still uses MDS as info system, so it failed as follows:

chown: failed to get attributes of `/opt/bdii/etc': No such file or directory
chown: failed to get attributes of `/opt/bdii/var': No such file or directory
ls: /usr/share/doc/openldap-servers-*: No such file or directory
/opt/glite/yaim/bin/../libexec/configure_node: line 12: [: : integer expression expected
sed: can't read /opt/bdii/etc/schemas: No such file or directory
sed: can't read /opt/bdii/etc/bdii.conf: No such file or directory
error reading information on service bdii: No such file or directory
bdii: unrecognized service
   INFO: Executing function: config_bdii_only 
bdii: unrecognized service

top BDII

• slapd could not be started, in bdii log, it complains :

/opt/glue/schema/openldap-2.1/Glue-CORE.schema: line 112: MatchingRule not found: "integerOrderingMatch"

so it seems in config_bdii, it's wrong to detect openldap version (actually no detection)

MON

• when calling config_rgma_server, it complains:

chown: failed to get attributes of `/var/log/glite/rgma-server/': No such file or directory
chmod: failed to get attributes of `/var/log/glite/rgma-server/': No such file or directory

it disappear when running yaim again.

• when calling config_rgma_gin, it complains "ERROR: SITE_BDII_HOST is not set", add it in site-info.def example and put in release notes?

VOBOX

• when calling config_vobox

ls: /opt/vobox/dteam/proxy_repository: Permission denied
 Starting ProxyRenewal Daemon: vobox-renewd                [  OK  ]
 Process  does not exist ...                               [FAILED]
ls: /opt/vobox/dteam/proxy_repository: Permission denied
 Starting ProxyRenewal Daemon: vobox-renewd                [  OK  ]
 Process  does not exist ...                               [FAILED]
ls: /opt/vobox/dteam/proxy_repository: Permission denied
 Starting ProxyRenewal Daemon: vobox-renewd                [  OK  ]
 Process  does not exist ...                               [FAILED]
ls: /opt/vobox/dteam/proxy_repository: Permission denied
 Starting ProxyRenewal Daemon: vobox-renewd                [  OK  ]
.......

and it also looks it tried to configue VOBOX for all sgm pool accounts, is it correct? And also it keeps printing out above error message on console after configuring. Notice: in grid-mapfile all sgm DN are mapped to sgm001, so we don't need to have daemons or cron for all sgm pool accounts.

classic SE

• when calling config_classicse, it complains:

/opt/glite/yaim/bin/../libexec/configure_node: line 5: +requires: command not found

Upgrade testing

Savannah patches

3.0 APT repository

3.1 YUM repository

External Sites testing reports

CERN testing report