LCG Grid Deployment - gLite Pre Production Services - Pre Production Coordination

Excerpts of this report are automatically published in OpsMeetingGliteReleases, OpsMeetingPilots, ScmPps

PPS Coordination Work Log


PPS Coordination Report on 25-Mar-10

EGEE Pre-Production Service Coordination:


2010-03-24: gLite 3.2 Update09 delivered to production
-- Release notes (preview) --

-- Main changes introduced --
  • (NEW) glite-LSF_utils for SL5
  • glite-GLEXEC_wn, glite-CREAM
    • New version of LCMAPS pep-c client supporting client authentication against the Argus framework.
  • LFC and DPM (new version 1.7.3 of client and server nodes) fixing in particular
    • Bug 53568, which could cause the DPM's SRM 2.2 server to crash
    • Bug 50686, which meant a malformed request could crash the LFC or the DPNS
    • The xrootd plugin for ALICE for the DPM has been upgraded to version 2.1.0.in order to increase the stability of the service.
    • WARNING: There are known issues upgrading from DPM 1.6.x (please check service release notes)
  • BDII:
    • bug fixes
    • (NEW) introduction of glue 2.0 info provider. The BDII noode starts being published according to the new schema
  • UI, WN, VOBOX
    • some rpms are removed to fix vulnerability issue for bug #54623
-- Early Adopter Sites for this update and reports -- Tentative date for release to production --
  • 22-Mar-10
-- Affected services currently NOT COVERED by the staged roll-out -- Please contact grid-deployment-managers@cern.ch to volunteer



PPS Coordination Report on 18-Mar-10

EGEE Pre-Production Service Coordination:


17-Feb-2010: Pilot of glexec/Argus in progress
check-point meeting held

  • Installation at CNAF T1 finished
  • Decision to use the production repository for future operations
  • Testing of the OSCT global banning list approved.
  • Pilot end date shifts to the 16th of April
  • Further developments and tests to be followed within the GDB
people interested in the activity are welcome to contact egee-pilot-argus@cern.ch .
Minutes of the meeting: https://twiki.cern.ch/twiki/bin/view/LCG/PPIslandKickOff2010x02x17
Pilot Home page: https://twiki.cern.ch/twiki/bin/view/EGEE/PilotServiceArgus

2010-03-16: gLite 3.1 Update 62 went to production
-- Release notes (preview) --

-- Main changes introduced --
  • glite-LB
    • extensions to LB notifications to allow export of LB data into MSG via LB harvester.
    • Due to the minimalistic backport of the functionality, the implementation is not optimal wrt. LB 2.0. This patch is provided as a quick solution, sites willing to run MSG export in production should migrate to LB 2.0 ASAP.
  • glite-SCAS
    • fixing bug #52648, which prevents user banning for the SCAS server
    • stability of the service improved significantly
    • several optimisations (see service release notes)
  • LCAS
    • Solved segmentation faults when a malformed proxy was provided by the calling library or application.
  • LFC and DPM (new version 1.7.3 of client and server nodes) fixing in particular
    • Bug 53568, which could cause the DPM's SRM 2.2 server to crash
    • Bug 50686, which meant a malformed request could crash the LFC or the DPNS
    • The xrootd plugin for ALICE for the DPM has been upgraded to version 2.1.0.in order to increase the stability of the service.
    • WARNING: There are known issues upgrading from DPM 1.6.x (please check service release notes)
  • APEL (affecting monbox and utils packages for various batch systems) with several bug fixes, in particular
    • APEL will now read the CPUScalingReferenceSI00 value from the site GIIS. If this value is not available, APEL will read GlueHostbenchmarkSI00.
    • SGE log files with format accounting-YYYYMMDD and accounting-YYYYMMDD.gz are now parsed.
    • Problems parsing different versions of LSF log files fixed. Also fixed parsing of parallel jobs in LSF.
    • APEL publisher can now use external public keys (from a file) to encrypt UserDNs.
  • glite-WMS
    • consolidation of current functionality for admins and users.
  • glite-UI, glite-WN, glite-VOBOX
    • New version of lcg-ManageVOTag solving the bug #52878
  • gLite-MPI-utils (backporting of fixes already deployed on SL5)
    • fix for a long known dependency issue with an incompatible torque library provided by the torque RPM.
    • minor bug fixes yaim configuration
    • default MPI installation is now OPENMPI, which is provided by the Scientific Linux distribution.
    • Support for MPICH1 and MPICH2 has been discontinued on SL4. MPICH2 is available for SL5.
    • KNOWN ISSUE: bug #62612 (missing version of MPI flavour in site BDII ) is NOT fixed in this version
-- Tentative date for release to production --
  • 15-Mar-10
-- Report on Staged Roll-out --
No major issues were reported by the sites who have applied this update.
  • A comment from GRIF about the DPM still running and old version of xrootd (the new one will be available with DPM 1.7.4)
  • More feedback from production sites would be desirable for the changes in the LB (glite-lb_R_1_9_4_2). Contacts have been made with CERN-PROD to try out the beta version later on this week.
  • All reports are available at http://www.cern.ch/pps/index.php?dir=./ActivityManagement/REPORTS/update=3.1.0_Bundle_05
  • Some affected services were NOT COVERED by any sites for the staged roll-out, namely
    • glite-CONDOR_utils_SL4_i386
    • glite-FTS_oracle_SL4_x86_64
    • glite-LFC_mysql_SL4_i386
    • glite-LFC_oracle_SL4_i386
    • glite-LFC_oracle_SL4_x86_64
    • glite-MPI_utils_SL4_i386
    • glite-SCAS_SL4_i386
    • glite-SE_dcache_admin_gdbm_SL4_x86_64
    • glite-SE_dcache_admin_postgres_SL4_x86_64
    • glite-SE_dcache_info_SL4_x86_64
    • glite-SE_dpm_disk_SL4_i386
    • glite-SE_dpm_disk_SL4_x86_64
    • glite-UI_SL4_x86_64
    • glite-VOBOX_SL4_i386
    • glite-VOBOX_SL4_x86_64
  • Some appointed Early Adopter sites couldn't perform some staged roll-out tasks for various reasons. Therefore operations report couldn't be collected for the following services
    • glite-LB_SL4_i386
    • glite-MON_SL4_i386
    • glite-SGE_utils_SL4_i386
    • glite-WN_SL4_i386
    • glite-WN_SL4_x86_64
    • IC-Info_registry_SL4_i386
Please contact grid-deployment-managers@cern.ch to volunteer



PPS Coordination Report on 11-Mar-10

EGEE Pre-Production Service Coordination:


2010-03-10: After staged roll-out the release of gLite 3.1 Update 62 is in preparation
-- Release notes (preview) --

-- Main changes introduced --
  • glite-LB
    • extensions to LB notifications to allow export of LB data into MSG via LB harvester.
    • Due to the minimalistic backport of the functionality, the implementation is not optimal wrt. LB 2.0. This patch is provided as a quick solution, sites willing to run MSG export in production should migrate to LB 2.0 ASAP.
  • glite-SCAS
    • fixing bug #52648, which prevents user banning for the SCAS server
    • stability of the service improved significantly
    • several optimisations (see service release notes)
  • LCAS
    • Solved segmentation faults when a malformed proxy was provided by the calling library or application.
  • LFC and DPM (new version 1.7.3 of client and server nodes) fixing in particular
    • Bug 53568, which could cause the DPM's SRM 2.2 server to crash
    • Bug 50686, which meant a malformed request could crash the LFC or the DPNS
    • The xrootd plugin for ALICE for the DPM has been upgraded to version 2.1.0.in order to increase the stability of the service.
    • WARNING: There are known issues upgrading from DPM 1.6.x (please check service release notes)
  • APEL (affecting monbox and utils packages for various batch systems) with several bug fixes, in particular
    • APEL will now read the CPUScalingReferenceSI00 value from the site GIIS. If this value is not available, APEL will read GlueHostbenchmarkSI00.
    • SGE log files with format accounting-YYYYMMDD and accounting-YYYYMMDD.gz are now parsed.
    • Problems parsing different versions of LSF log files fixed. Also fixed parsing of parallel jobs in LSF.
    • APEL publisher can now use external public keys (from a file) to encrypt UserDNs.
  • glite-WMS
    • consolidation of current functionality for admins and users.
  • glite-UI, glite-WN, glite-VOBOX
    • New version of lcg-ManageVOTag solving the bug #52878
  • gLite-MPI-utils (backporting of fixes already deployed on SL5)
    • fix for a long known dependency issue with an incompatible torque library provided by the torque RPM.
    • minor bug fixes yaim configuration
    • default MPI installation is now OPENMPI, which is provided by the Scientific Linux distribution.
    • Support for MPICH1 and MPICH2 has been discontinued on SL4. MPICH2 is available for SL5.
    • KNOWN ISSUE: bug #62612 (missing version of MPI flavour in site BDII ) is NOT fixed in this version
-- Tentative date for release to production --
  • 15-Mar-10
-- Report on Staged Roll-out --
No major issues were reported by the sites who have applied this update.
  • A comment from GRIF about the DPM still running and old version of xrootd (the new one will be available with DPM 1.7.4)
  • More feedback from production sites would be desirable for the changes in the LB (glite-lb_R_1_9_4_2). Contacts have been made with CERN-PROD to try out the beta version later on this week.
  • All reports are available at http://www.cern.ch/pps/index.php?dir=./ActivityManagement/REPORTS/update=3.1.0_Bundle_05
  • Some affected services were NOT COVERED by any sites for the staged roll-out, namely
    • glite-CONDOR_utils_SL4_i386
    • glite-FTS_oracle_SL4_x86_64
    • glite-LFC_mysql_SL4_i386
    • glite-LFC_oracle_SL4_i386
    • glite-LFC_oracle_SL4_x86_64
    • glite-MPI_utils_SL4_i386
    • glite-SCAS_SL4_i386
    • glite-SE_dcache_admin_gdbm_SL4_x86_64
    • glite-SE_dcache_admin_postgres_SL4_x86_64
    • glite-SE_dcache_info_SL4_x86_64
    • glite-SE_dpm_disk_SL4_i386
    • glite-SE_dpm_disk_SL4_x86_64
    • glite-UI_SL4_x86_64
    • glite-VOBOX_SL4_i386
    • glite-VOBOX_SL4_x86_64
  • Some appointed Early Adopter sites couldn't perform some staged roll-out tasks for various reasons. Therefore operations report couldn't be collected for the following services
    • glite-LB_SL4_i386
    • glite-MON_SL4_i386
    • glite-SGE_utils_SL4_i386
    • glite-WN_SL4_i386
    • glite-WN_SL4_x86_64
    • IC-Info_registry_SL4_i386
Please contact grid-deployment-managers@cern.ch to volunteer

2010-03-09: Staged roll-out of several updates to gLite 3.2 starting
-- Release notes (preview) --

-- Main changes introduced --
  • (NEW) glite-LSF_utils for SL5
  • glite-GLEXEC_wn, glite-CREAM
    • New version of LCMAPS pep-c client supporting client authentication against the Argus framework.
  • LFC and DPM (new version 1.7.3 of client and server nodes) fixing in particular
    • Bug 53568, which could cause the DPM's SRM 2.2 server to crash
    • Bug 50686, which meant a malformed request could crash the LFC or the DPNS
    • The xrootd plugin for ALICE for the DPM has been upgraded to version 2.1.0.in order to increase the stability of the service.
    • WARNING: There are known issues upgrading from DPM 1.6.x (please check service release notes)
    • (NEW) introduction of glue 2.0 info provider
  • BDII: bug fixes
  • UI, WN, VOBOX
    • some rpms are removed to fix vulnerability issue for bug #54623
-- Early Adopter Sites for this update and reports -- Tentative date for release to production --
  • 22-Mar-10
-- Affected services currently NOT COVERED by the staged roll-out -- Please contact grid-deployment-managers@cern.ch to volunteer


PPS Coordination Report on 4-Mar-10

EGEE Pre-Production Service Coordination:


2010-03-02: Staged roll-out of several updates to gLite 3.1 starting
-- Release notes (preview) --

-- Main changes introduced --
  • glite-LB
    • extensions to LB notifications to allow export of LB data into MSG via LB harvester.
    • Due to the minimalistic backport of the functionality, the implementation is not optimal wrt. LB 2.0. This patch is provided as a quick solution, sites willing to run MSG export in production should migrate to LB 2.0 ASAP.
  • glite-SCAS
    • fixing bug #52648, which prevents user banning for the SCAS server
    • stability of the service improved significantly
    • several optimisations (see service release notes)
  • LCAS
    • Solved segmentation faults when a malformed proxy was provided by the calling library or application.
  • LFC and DPM (new version 1.7.3 of client and server nodes) fixing in particular
    • Bug 53568, which could cause the DPM's SRM 2.2 server to crash
    • Bug 50686, which meant a malformed request could crash the LFC or the DPNS
    • The xrootd plugin for ALICE for the DPM has been upgraded to version 2.1.0.in order to increase the stability of the service.
    • WARNING: There are known issues upgrading from DPM 1.6.x (please check service release notes)
  • APEL (affecting monbox and utils packages for various batch systems) with several bug fixes, in particular
    • APEL will now read the CPUScalingReferenceSI00 value from the site GIIS. If this value is not available, APEL will read GlueHostbenchmarkSI00.
    • SGE log files with format accounting-YYYYMMDD and accounting-YYYYMMDD.gz are now parsed.
    • Problems parsing different versions of LSF log files fixed. Also fixed parsing of parallel jobs in LSF.
    • APEL publisher can now use external public keys (from a file) to encrypt UserDNs.
  • glite-WMS
    • consolidation of current functionality for admins and users.
  • glite-UI, glite-WN, glite-VOBOX
    • New version of lcg-ManageVOTag solving the bug #52878
  • gLite-MPI-utils (backporting of fixes already deployed on SL5)
    • fix for a long known dependency issue with an incompatible torque library provided by the torque RPM.
    • minor bug fixes yaim configuration
    • default MPI installation is now OPENMPI, which is provided by the Scientific Linux distribution.
    • Support for MPICH1 and MPICH2 has been discontinued on SL4. MPICH2 is available for SL5.
    • KNOWN ISSUE: bug #62612 (missing version of MPI flavour in site BDII ) is NOT fixed in this version
-- Early Adopter Sites for this update and reports -- Tentative date for release to production --
  • 10-Mar-10
-- Affected services currently NOT COVERED by the staged roll-out --
  • glite-CONDOR_utils_SL4_i386
  • glite-FTS_oracle_SL4_x86_64
  • glite-LFC_mysql_SL4_i386
  • glite-LFC_oracle_SL4_i386
  • glite-LFC_oracle_SL4_x86_64
  • glite-MPI_utils_SL4_i386
  • glite-SCAS_SL4_i386
  • glite-SE_dcache_admin_gdbm_SL4_x86_64
  • glite-SE_dcache_admin_postgres_SL4_x86_64
  • glite-SE_dcache_info_SL4_x86_64
  • glite-SE_dpm_disk_SL4_i386
  • glite-SE_dpm_disk_SL4_x86_64
  • glite-UI_SL4_x86_64
  • glite-VOBOX_SL4_i386
  • glite-VOBOX_SL4_x86_64
Please contact grid-deployment-managers@cern.ch to volunteer

2010-02-18: gLite 3.1 Update 61 released to production
-- Release notes (preview) --

-- Main changes introduced --
  • NEW gLExec tests to glite-swat-client
  • WMS
    • bug fixes for proxy renewal mechanism
    • bug fix 58347: access to catalog via DLI
  • FTS 2.2.3
    • Fix for bug #59955: FTS: agent crash when picking up a job (glite-data-agents-common v1.1.1-1 2009-12-08)
    • Fix for bug #60095: FTS: Couldn't set the private key (glite-data-transfer-fts v3.7.0-3 2010-01-18)
    • Updates for more liberal checksum handling and to relax requirements on how long an SRM TURL is kept valid.
    • Finalization of the checksum support.
    • The database schema has changed, one has to follow the instructions of the Yaim script to do the upgrade.
    • The interface between the transfer-agents and transfer-url-copy has changed, so one has to stop and drain the channels before doing an upgrade. * FTA
    • bug fixes for proxy renewal mechanism
    • dcap 1.9.3-5
    • dcache-srmclient 1.9.3-5
    • dcap clients re-packaged
  • SGE utils
    • support for "GlueCEPolicyAssignedJobSlots"
  • YAIM core 4.0.10 and 4.0.11
    • new function to configure the new myproxy server and clients
    • configuration of new SAGA gLite adapters.
    • grid environment cleaning mechanism to be able to install different WN tarball versions in the same machine.
    • bug fixes
  • YAIM clients
    • improves tarball installation of WNs
  • Update to the SLCS client (Short Lived Credential Service)
  • New version of lcg-vomscerts
    • next cert for lcg-voms.cern.ch and removes old certs for cclcgvomsli01.in2p3.fr and vo.racf.bnl.gov
-- Early Adopter Sites for this update and reports



PPS Coordination Report on 18-Feb-10

EGEE Pre-Production Service Coordination:


2010-03-01: Staged roll-out of several updates to gLite 3.1 starting
-- Release notes (preview) --

-- Main changes introduced --
  • NEW gLExec tests to glite-swat-client
  • WMS
    • bug fixes for proxy renewal mechanism
    • bug fix 58347: access to catalog via DLI
  • FTS 2.2.3
    • Fix for bug #59955: FTS: agent crash when picking up a job (glite-data-agents-common v1.1.1-1 2009-12-08)
    • Fix for bug #60095: FTS: Couldn't set the private key (glite-data-transfer-fts v3.7.0-3 2010-01-18)
    • Updates for more liberal checksum handling and to relax requirements on how long an SRM TURL is kept valid.
    • Finalization of the checksum support.
    • The database schema has changed, one has to follow the instructions of the Yaim script to do the upgrade.
    • The interface between the transfer-agents and transfer-url-copy has changed, so one has to stop and drain the channels before doing an upgrade. * FTA
    • bug fixes for proxy renewal mechanism
    • dcap 1.9.3-5
    • dcache-srmclient 1.9.3-5
    • dcap clients re-packaged
  • SGE utils
    • support for "GlueCEPolicyAssignedJobSlots"
  • YAIM core 4.0.10 and 4.0.11
    • new function to configure the new myproxy server and clients
    • configuration of new SAGA gLite adapters.
    • grid environment cleaning mechanism to be able to install different WN tarball versions in the same machine.
    • bug fixes
  • YAIM clients
    • improves tarball installation of WNs
  • Update to the SLCS client (Short Lived Credential Service)
-- Early Adopter Sites for this update and reports -- Tentative date for release to production --
  • 16-Feb-10
-- Affected services currently NOT COVERED by the staged roll-out --
  • glite-AMGA_postgres_SL4_i386
  • glite-CONDOR_utils_SL4_i386
  • glite-FTA_oracle_SL4_x86_64
  • glite-FTM_SL4_i386
  • glite-FTM_SL4_x86_64
  • glite-FTS_oracle_SL4_x86_64
  • glite-HYDRA_mysql_SL4_i386
  • glite-LFC_mysql_SL4_i386
  • glite-LFC_oracle_SL4_i386
  • glite-LFC_oracle_SL4_x86_64
  • glite-MPI_utils_SL4_i386
  • glite-PX_SL4_i386
  • glite-SCAS_SL4_i386
  • glite-SE_dpm_disk_SL4_i386
  • glite-SE_dpm_disk_SL4_x86_64
  • glite-SLCS_client_SL4_i386
  • glite-VOBOX_SL4_i386
Please contact grid-deployment-managers@cern.ch to volunteer

2010-02-15: Staged roll-out of several updates to gLite 3.1 in progress
NOTE: the release to production has to be delayed by one day with respect to what originally planned due to a deployment issue affecting the dcache server and the analysis of an issue affecting the UI of one of the early adopter sites. The dcache server has been removed from the release and a new round of testing is now ongoing to validate the correct interactions of the clients with the existing service.
-- Release notes (preview) --

-- Main changes introduced --
  • NEW gLExec tests to glite-swat-client
  • WMS
    • bug fixes for proxy renewal mechanism
    • bug fix 58347: access to catalog via DLI
  • FTS 2.2.3
    • Fix for bug #59955: FTS: agent crash when picking up a job (glite-data-agents-common v1.1.1-1 2009-12-08)
    • Fix for bug #60095: FTS: Couldn't set the private key (glite-data-transfer-fts v3.7.0-3 2010-01-18)
    • Updates for more liberal checksum handling and to relax requirements on how long an SRM TURL is kept valid.
    • Finalization of the checksum support.
    • The database schema has changed, one has to follow the instructions of the Yaim script to do the upgrade.
    • The interface between the transfer-agents and transfer-url-copy has changed, so one has to stop and drain the channels before doing an upgrade. * FTA
    • bug fixes for proxy renewal mechanism
    • dcap 1.9.3-5
    • dcache-srmclient 1.9.3-5
    • dcap clients re-packaged
  • SGE utils
    • support for "GlueCEPolicyAssignedJobSlots"
  • YAIM core 4.0.10 and 4.0.11
    • new function to configure the new myproxy server and clients
    • configuration of new SAGA gLite adapters.
    • grid environment cleaning mechanism to be able to install different WN tarball versions in the same machine.
    • bug fixes
  • YAIM clients
    • improves tarball installation of WNs
  • Update to the SLCS client (Short Lived Credential Service)
-- Early Adopter Sites for this update and reports -- Tentative date for release to production --
  • 16-Feb-10
-- Affected services currently NOT COVERED by the staged roll-out --
  • glite-AMGA_postgres_SL4_i386
  • glite-CONDOR_utils_SL4_i386
  • glite-FTA_oracle_SL4_x86_64
  • glite-FTM_SL4_i386
  • glite-FTM_SL4_x86_64
  • glite-FTS_oracle_SL4_x86_64
  • glite-HYDRA_mysql_SL4_i386
  • glite-LFC_mysql_SL4_i386
  • glite-LFC_oracle_SL4_i386
  • glite-LFC_oracle_SL4_x86_64
  • glite-MPI_utils_SL4_i386
  • glite-PX_SL4_i386
  • glite-SCAS_SL4_i386
  • glite-SE_dpm_disk_SL4_i386
  • glite-SE_dpm_disk_SL4_x86_64
  • glite-SLCS_client_SL4_i386
  • glite-VOBOX_SL4_i386
Please contact grid-deployment-managers@cern.ch to volunteer



PPS Coordination Report on 11-Feb-10

EGEE Pre-Production Service Coordination:


2010-02-8: gLite 3.2 Update 08 was released to production
-- Release notes -- -- Main changes introduced --
  • NEW gLExec tests to glite-swat-client
  • NEW Argus service for SL5
  • NEW glite-SE_dpm_disk intiial release for SL5
  • dcache
    • dcap 1.9.3-5 dcache-srmclient 1.9.5-3 for SL5
    • dcap clients re-packaged
  • new version of lcg-infosites version solving the bug #37572
  • gLite-MPI-utils
    • fix for a long known dependency issue with an incompatible torque library provided by the torque RPM.
    • minor bug fixes yaim configuration
    • default MPI installation is now OPENMPI, which is provided by the Scientific Linux distribution.
-- Reports from early adopters sites --



EGEE-III PPS Quarterly Report Q7 for UE

Preproduction Service
The decommissioning of the PPS infrastructure is in progress since October 09 with the activation of the staged roll-out process. As programmed there were no more releases deployed through PPS since then. Some of the grid and operational services of the PPS (e.g. top level BDIIs, PPS registry, SAM) are still maintained because they are used as support tool during the hand over to EGI. A number of former PPS sites have started a process of reintegration of their resources within the corresponding production sites. This is a slow and heterogeneous process as very influenced by local site policies. Therefore the task coordinator is essentially treating the process case by case.
A new pilot service focused on the Argus authorisation service was started in November with the participation of 6 production sites (CESNET, KIT, INFN-CNAF,INFN-T1, SRCE, SWITCH). Three of the four LHCb experiments (Alice, CMS and Alice) have accepted to be involved in the operations at various levels focusing on the use of the glexec capability. The tentavie end-date for this pilot is set to the 31st of March 2010.
In the last quarter the new responsible people of the EGI task O.E.9 (Coordination of middleware roll-out and deployment) will be associated in the coordination of the relevant functions as well as in the operations of the support tools. This tutoring period should last until the end of the project. The last technical integration activity (support in the GOCDB for Early adopter sites registration) will be finalised in QE8 as well.

Middleware Deployment
Mainly due to the start of the LHC operations a natural slow down of the releases of gLite updates to the production system was observed. In fact only two updates were released to production in this quarter on the two active baselines. Both updates were released using the 'staged roll-out' approach. Particularly welcome has been the release of the SL5 version of the glexec/SCAS identity switching capabilty and of the CREAM CE. CREAM is now in production at all effects, being monitored by the Operators on Duty as a critical service. Availability and reliability algorithms though still need to be updated as they are still computed looking only at the lcg-CE.
A list of the recommended middleware versions for the production service as well as the process to maintain it has been compiled in collaboration with the WLCG. The lists of supported client and server versions are available respectively at https://twiki.cern.ch/twiki/bin/view/EGEE/SupportedClientVersions and https://twiki.cern.ch/twiki/bin/view/EGEE/SupportedServiceVersions .
Plans for the next quarter include the enhancement of the staged roll-out process by affiliation of more early adopter sites



PPS Coordination Report on 4-Feb-10

EGEE Pre-Production Service Coordination:


2010-02-03: Staged roll-out of several updates to gLite 3.1 in progress
-- Release notes (preview) --

-- Main changes introduced --
  • NEW gLExec tests to glite-swat-client
  • WMS
    • bug fixes for proxy renewal mechanism
    • bug fix 58347: access to catalog via DLI
  • FTS 2.2.1
    • Checksum support.
    • Database schema changed, follow the instructions of the Yaim script to upgrade.
    • The interface between the transfer-agents and transfer-url-copy has changed, need to stop and drain the channels before doing an upgrade.
  • FTA
    • bug fixes for proxy renewal mechanism
  • dcache
    • dcache-server-1.9.1-11 (security update)
    • dcap 1.9.3-5
    • dcache-srmclient 1.9.5-3
    • dcap clients re-packaged
  • SGE utils
    • support for "GlueCEPolicyAssignedJobSlots"
  • YAIM core 4.0.10 and 4.0.11
    • new function to configure the new myproxy server and clients
    • configuration of new SAGA gLite adapters.
    • grid environment cleaning mechanism to be able to install different WN tarball versions in the same machine.
    • bug fixes
  • YAIM clients
    • improves tarball installation of WNs
  • Update to the SLCS client (Short Lived Credential Service)
-- Early Adopter Sites for this update and reports -- Tentative date for release to production --
  • 15-Feb-10 ( to be confirmed )
-- Affected services currently NOT COVERED by the staged roll-out --
  • glite-AMGA_postgres_SL4_i386
  • glite-CONDOR_utils_SL4_i386
  • glite-FTA_oracle_SL4_x86_64
  • glite-FTM_SL4_i386
  • glite-FTM_SL4_x86_64
  • glite-FTS_oracle_SL4_x86_64
  • glite-HYDRA_mysql_SL4_i386
  • glite-LFC_mysql_SL4_i386
  • glite-LFC_oracle_SL4_i386
  • glite-LFC_oracle_SL4_x86_64
  • glite-MPI_utils_SL4_i386
  • glite-PX_SL4_i386
  • glite-SCAS_SL4_i386
  • glite-SE_dcache_admin_gdbm_SL4_i386
  • glite-SE_dcache_admin_gdbm_SL4_x86_64
  • glite-SE_dcache_admin_postgres_SL4_i386
  • glite-SE_dcache_admin_postgres_SL4_x86_64
  • glite-SE_dcache_info_SL4_i386
  • glite-SE_dcache_info_SL4_x86_64
  • glite-SE_dcache_pool_SL4_i386
  • glite-SE_dcache_pool_SL4_x86_64
  • glite-SE_dpm_disk_SL4_i386
  • glite-SE_dpm_disk_SL4_x86_64
  • glite-SLCS_client_SL4_i386
  • glite-VOBOX_SL4_i386
Please contact grid-deployment-managers@cern.ch to volunteer

2010-02-03: release of gLite 3.2 Update 08 in preparation
-- Release notes (preview) --

-- Main changes introduced --
  • NEW gLExec tests to glite-swat-client
  • NEW Argus service for SL5
  • NEW glite-SE_dpm_disk intiial release for SL5
  • dcache
    • dcap 1.9.3-5 dcache-srmclient 1.9.5-3 for SL5
    • dcap clients re-packaged
  • new version of lcg-infosites version solving the bug #37572
  • gLite-MPI-utils
    • fix for a long known dependency issue with an incompatible torque library provided by the torque RPM.
    • minor bug fixes yaim configuration
    • default MPI installation is now OPENMPI, which is provided by the Scientific Linux distribution.
-- Reports from early adopters sites -- -- Tentative date for release to production --
  • 8-Feb-10 ( still under discussion )

2-Feb-2010: Pilot of glexec/Argus in progress
check-point meeting held

  • All sites will be soon requested to upgrade to the new version of Argus PATCH:3536 . CANF-T1 will be the first, the other will follow
  • All sites requested to apply the workaround in BUG:62206 in order for the Argus servers to star being published in the information system.
  • Integration works in progress for Alice
  • Integration works confirmed to start at mid February for CMS
Sites interested in the activity are welcome to contact egee-pilot-argus@cern.ch .
Minutes of the meeting: https://twiki.cern.ch/twiki/bin/view/LCG/PPIslandKickOff2010x02x02
Pilot Home page: https://twiki.cern.ch/twiki/bin/view/EGEE/PilotServiceArgus


PPS Coordination Report on 28-Jan-10

EGEE Pre-Production Service Coordination:


25-Jan-10: Sites willing to subscribe for the staged roll-out of the following services are needed for the new year

  • glite-ARGUS on SL5
  • glite-LFC_oracle on SL5
  • glite-MPI_utils on SL5

2010-01-18: Staged roll-out of several updates to gLite 3.2 in progress
Some installation issues were found affecting the dcache/dcap client packages. Unfortunately these cannot be simply removed from the release because they are wrapped in the client nodes (UI and WN). There is an analysis in progress to verify the impact of the issue and find a suitable workaround. https://gus.fzk.de/ws/ticket_info.php?ticket=53922
-- Release notes (preview) --

-- Main changes introduced --
  • NEW gLExec tests to glite-swat-client
  • NEW Argus service for SL5
  • NEW glite-SE_dpm_disk intiial release for SL5
  • dcache
    • dcap 1.9.3-5 dcache-srmclient 1.9.5-3 for SL5
    • dcap clients re-packaged
  • new version of lcg-infosites version solving the bug #37572
  • gLite-MPI-utils
    • fix for a long known dependency issue with an incompatible torque library provided by the torque RPM.
    • minor bug fixes yaim configuration
    • default MPI installation is now OPENMPI, which is provided by the Scientific Linux distribution.
-- Early Adopter Sites for this update --
  • CERN_PPS: (afs_UI, Top BDII, WN_glexec, SL5, x86_64)
  • AEGIS01-IPB-SCL (glite-CREAM_TORQUE, SL5, x86_64)
  • SAMPA (glite-CREAM_TORQUE, SL5, x86_64)
  • INFN-PADOVA (glite-LFC_mysql, SL5, x86_64)
  • ITWM-PPS (glite-SE_dpm_disk/mysql, SL5, x86_64)
  • DESY-PPS (glite-LFC_mysql, glite-VOBOX, SL5, x86_64)
  • RAL-LCG2 (Top BDII, SL5, x86_64)
  • FZK (WN_GLEXEC, (SCAS) SL5, x86_64)
-- Reports -- -- Tentative date for release to production --
  • 27-Jan-10 ( still under discussion )
-- Affected services currently NOT COVERED by the staged roll-out --
  • glite-ARGUS_SL5_x86_64
  • glite-LFC_oracle_SL5_x86_64
  • glite-MPI_utils_SL5_x86_64
Please contact grid-deployment-managers@cern.ch to volunteer

12-Jan-2010: Pilot of glexec/Argus in progress
check-point meeting held

  • All the sites met the deadline of the 15th of January for the installation of the glexec/argus chain
  • additional requirement setr by Alice of a third VOBOX in order to start the tests (accepted by CNAF and KIT)
The following milestones were set
  • 1st site technically available for Experiments to test (SWITCH): 1-Dec
  • All sites technically available for Experiments to test: 15-Jan
  • Indicative start of Alice developments to integrate glexec: 18-Jan
  • Indicative start of CMS developments to integrate glexec: 15-Feb
  • END of activity (proposed): 31-Mar
Sites interested in taking part in the activity are welcome to contact egee-pilot-argus@cern.ch .
Minutes of the meeting: https://twiki.cern.ch/twiki/bin/view/LCG/PPIslandKickOff2010x01x12
Pilot Home page: https://twiki.cern.ch/twiki/bin/view/EGEE/PilotServiceArgus


PPS Coordination Report on 14-Jan-10

EGEE Pre-Production Service Coordination:


2010-1-07: Release of gLite 3.2 Update07 was released to production
-- Release notes --

-- Main changes introduced --
  • NEW TORQUE server client and utils for SL5
    • torque version: 2.3.6, maui version is 3.2.6p21 (several bug fixes)
    • maui configuration can be now disabled in YAIM
    • torque infoproviders now also populate GlueCEPolicyAssignedJobSlots for the GlueCE.
    • Note that torque server and client version have to be the same on a site for a proper interaction. Keep this in mind for the case of mixed SL4/SL5 torque client/server
  • NEW SCAS (Site Central Authorization Service) for SL5
  • NEW GLEXEC for SL5
  • NEW CREAM for SL5

2010-01-07: Release of gLite 3.1 Update 60 was released to production
-- Release notes --

-- Main changes introduced --
  • Update to TORQUE server, client and utils for SL4
    • torque version: 2.3.6, maui version is 3.2.6p21 (several bug fixes)
    • Fix for a security vulnerability
    • maui configuration can be now disabled in YAIM
    • torque infoproviders now also populate GlueCEPolicyAssignedJobSlots for the GlueCE.
    • Note that torque server and client version have to be the same on a site for a proper interaction. Keep this in mind for the case of mixed SL4/SL5 torque client/server
    • Known Issue : problems in deployment of MPI together with torque_client. Fix currently in certification
  • Update to WMS
    • New version of ICE for the WMS to fix bug #59054
  • Update to CREAM
    • Bug fixes
    • With respect to the previous version this one fixes a couple of problems with the values published in the information service (in particular it enables monitoring to "fix installed capacity")
    • A fix for a vulnerability affecting BLAH is included as well
  • New certificate for two voms servers



PPS Coordination Report on 7-Jan-10

EGEE Pre-Production Service Coordination:


2010-1-05: Release of gLite 3.2 Update07 to production in preparation
-- Release notes (preview) --

-- Main changes introduced --
  • NEW TORQUE server client and utils for SL5
    • torque version: 2.3.6, maui version is 3.2.6p21 (several bug fixes)
    • maui configuration can be now disabled in YAIM
    • torque infoproviders now also populate GlueCEPolicyAssignedJobSlots for the GlueCE.
    • Note that torque server and client version have to be the same on a site for a proper interaction. Keep this in mind for the case of mixed SL4/SL5 torque client/server
  • NEW SCAS (Site Central Authorization Service) for SL5
  • NEW GLEXEC for SL5
  • NEW CREAM for SL5
-- Early Adopter Sites for this update --
  • CERN_PPS: (afs_UI, WN_GLEXEC, SL5, x86_64)
  • SAMPA (glite-CREAM_TORQUE, SL5, x86_64)
  • DESY-PPS (glite-LFC_mysql glite-VOBOX, SL5, x86_64)
  • INFN-PADOVA (glite-LFC_mysql, SL5, x86_64)
  • ITWM-PPS (glite-SE_dpm_mysql, WN_GLEXEC, SL5, x86_64)
  • AEGIS01-IPB-SCL (glite-CREAM_TORQUE, SL5, x86_64)
  • FZK (WN_GLEXEC, SL5, x86_64)
-- Reports -- -- Tentative date for release to production --
  • 5-Jan-09 ( still under discussion )
-- Affected services currently NOT COVERED by the staged roll-out --
  • glite-SCAS (SL5, x86_64)
  • gLite-TORQUE server (SL5, x86_64)
Please contact grid-deployment-managers@cern.ch to volunteer

2010-01-06: Release of gLite 3.1 Update 60 to production in preparation
-- Release notes (preview) --

-- Main changes introduced --
  • Update to TORQUE server, client and utils for SL4
    • torque version: 2.3.6, maui version is 3.2.6p21 (several bug fixes)
    • Fix for a security vulnerability
    • maui configuration can be now disabled in YAIM
    • torque infoproviders now also populate GlueCEPolicyAssignedJobSlots for the GlueCE.
    • Note that torque server and client version have to be the same on a site for a proper interaction. Keep this in mind for the case of mixed SL4/SL5 torque client/server
    • Known Issue : problems in deployment of MPI together with torque_client. Fix currently in certification
  • Update to WMS
    • New version of ICE for the WMS to fix bug #59054
  • Update to CREAM
    • Bug fixes
    • With respect to the previous version this one fixes a couple of problems with the values published in the information service (in particular it enables monitoring to "fix installed capacity")
    • A fix for a vulnerability affecting BLAH is included as well
  • New certificate for two voms servers
-- Early Adopter Sites for this update --
  • SiGNET (glite-CREAM_TORQUE, lcg-CE_TORQUE SL4, i386)
  • PPS-IFIC (glite-WMS, SL3, i386)
  • ITWM-PPS (glite-SE_dpm_mysql, WN_GLEXEC, SL4, x86_64)
  • WCSS-PPS (lcg-CE_TORQUE, SL44 x86_64
  • GUP-CERTIF-TB (lcg-CE_TORQUE, SL4, i386)
-- Reports -- -- Tentative date for release to production --
  • 6-Jan-09
-- Affected services currently NOT COVERED by the staged roll-out --
  • glite-MPI_utils (SL4, i386)



PPS Coordination Report on 10-Dec-09

EGEE Pre-Production Service Coordination:


7-Dec-09: Sites willing to subscribe for the staged roll-out of the following services are needed for the new year

  • FTS on SL4 and SL5
  • Site BDII on SL4 and SL5
  • Top-level BDII on SL4 and SL4

2009-12-07: Staged roll-out of several updates to gLite 3.2 in progress
Some issues were find a one of the sites (currently under investigation) https://gus.fzk.de/ws/ticket_info.php?ticket=53922

-- Release notes (preview) --

-- Main changes introduced --
  • NEW TORQUE server client and utils for SL5
    • torque version: 2.3.6, maui version is 3.2.6p21 (several bug fixes)
    • maui configuration can be now disabled in YAIM
    • torque infoproviders now also populate GlueCEPolicyAssignedJobSlots for the GlueCE.
    • Note that torque server and client version have to be the same on a site for a proper interaction. Keep this in mind for the case of mixed SL4/SL5 torque client/server
  • NEW SCAS (Site Central Authorization Service) for SL5
  • NEW GLEXEC for SL5
  • NEW CREAM for SL5
-- Early Adopter Sites for this update --
  • CERN_PPS: (afs_UI, WN_GLEXEC, SL5, x86_64)
  • SAMPA (glite-CREAM_TORQUE, SL5, x86_64)
  • DESY-PPS (glite-LFC_mysql, SL5, x86_64)
  • INFN-PADOVA (glite-LFC_mysql, SL5, x86_64)
  • ITWM-PPS (glite-SE_dpm_mysql, WN_GLEXEC, SL5, x86_64)
  • AEGIS01-IPB-SCL (glite-CREAM_TORQUE, SL5, x86_64)
  • FZK (WN_GLEXEC, SL5, x86_64)
-- Reports -- -- Tentative date for release to production --
  • 5-Jan-09 ( still under discussion )
-- Affected services currently NOT COVERED by the staged roll-out --
  • glite-SCAS (SL5, x86_64)
  • glite-VOBOX (SL5, x86_64)
  • gLite-TORQUE server (SL%, x86_64)
Please contact grid-deployment-managers@cern.ch to volunteer

2009-12-07: Staged roll-out of several updates to gLite 3.1 in progress
-- Release notes (preview) --

-- Main changes introduced --
  • Update to TORQUE server, client and utils for SL4
    • torque version: 2.3.6, maui version is 3.2.6p21 (several bug fixes)
    • Fix for a security vulnerability
    • maui configuration can be now disabled in YAIM
    • torque infoproviders now also populate GlueCEPolicyAssignedJobSlots for the GlueCE.
    • Note that torque server and client version have to be the same on a site for a proper interaction. Keep this in mind for the case of mixed SL4/SL5 torque client/server
    • Known Issue : problems in deployment of MPI together with torque_client. Fix currently in certification
  • Update to WMS
    • New version of ICE for the WMS to fix bug #59054
  • Update to CREAM
    • Bug fixes
    • With respect to the previous version this one fixes a couple of problems with the values published in the information service (in particular it enables monitoring to "fix installed capacity")
    • A fix for a vulnerability affecting BLAH is included as well
-- Early Adopter Sites for this update --
  • SiGNET (glite-CREAM_TORQUE, lcg-CE_TORQUE SL4, i386)
  • PPS-IFIC (glite-WMS, SL3, i386)
  • ITWM-PPS (glite-SE_dpm_mysql, WN_GLEXEC, SL4, x86_64)
  • WCSS-PPS (lcg-CE_TORQUE, SL44 x86_64
  • GUP-CERTIF-TB (lcg-CE_TORQUE, SL4, i386)
-- Reports -- -- Tentative date for release to production --
  • 5-Jan-09 ( still under discussion )
-- Affected services currently NOT COVERED by the staged roll-out --
  • glite-MPI_utils (SL4, i386)
Please contact grid-deployment-managers@cern.ch to volunteer



PPS Coordination Report on 3-Dec-09

EGEE Pre-Production Service Coordination:


1-Dec-2009: Pilot of glexec/Argus in progress
A meeting with the experiments was held in order to complete the planning for the pilot
A kick-off meeting with 5 interested sites was held today and an initial timeline was defined.
The following milestones were set

  • 1st site technically available for Experiments to test (SWITCH): 1-Dec
  • All sites technically available for Experiments to test: 15-Jan
  • Indicative start of Alice developments to integrate glexec: 18-Jan
  • Indicative start of CMS developments to integrate glexec: 15-Feb
  • END of activity (proposed): 31-Mar
Sites interested in taking part in the activity are welcome to contact egee-pilot-argus@cern.ch .
Minutes of the meeting: https://twiki.cern.ch/twiki/bin/view/LCG/PPIslandKickOff2009x12x01
Pilot Home page: https://twiki.cern.ch/twiki/bin/view/EGEE/PilotServiceArgus

30-Nov-09: Sites willing to subscribe for the staged roll-out of the following services are URGENTLY needed

  • glexec/SCAS on SL4 and SL5
  • CREAM on SL5
  • VOBOX on SL4 and SL5
  • gLite-TORQUE server on SL5

2009-11-30: Staged roll-out of several updates to gLite 3.2 started
-- Release notes (preview) --

-- Main changes introduced --
  • NEW TORQUE server client and utils for SL5
    • torque version: 2.3.6, maui version is 3.2.6p21 (several bug fixes)
    • maui configuration can be now disabled in YAIM
    • torque infoproviders now also populate GlueCEPolicyAssignedJobSlots for the GlueCE.
    • Note that torque server and client version have to be the same on a site for a proper interaction. Keep this in mind for the case of mixed SL4/SL5 torque client/server
  • NEW SCAS (Site Central Authorization Service) for SL5
  • NEW GLEXEC for SL5
  • NEW CREAM for SL5
-- Early Adopter Sites for this update --
  • CERN_PPS: (afs_UI, WN_GLEXEC, SL5, x86_64)
  • SAMPA (glite-CREAM_TORQUE, SL5, x86_64)
  • DESY-PPS (glite-LFC_mysql, SL5, x86_64)
  • INFN-PADOVA (glite-LFC_mysql, SL5, x86_64)
  • ITWM-PPS (glite-SE_dpm_mysql, WN_GLEXEC, SL5, x86_64)
-- Reports -- -- Tentative date for release to production --
  • 8-Dec-09
-- Affected services currently NOT COVERED by the staged roll-out --
  • glite-SCAS (SL5, x86_64)
  • glite-VOBOX (SL5, x86_64)
  • gLite-TORQUE server (SL%, x86_64)
Please contact grid-deployment-managers@cern.ch to volunteer

26-Nov-09: Sites willing to subscribe for the staged roll-out of the following services are needed

  • glite-MPI_utils (SL4, i386)
  • CREAM on SL4

2009-11-26: Staged roll-out of several updates to gLite 3.1 started
-- Release notes (preview) --

-- Main changes introduced --
  • Update to TORQUE server, client and utils for SL4
    • torque version: 2.3.6, maui version is 3.2.6p21 (several bug fixes)
    • Fix for a security vulnerability
    • maui configuration can be now disabled in YAIM
    • torque infoproviders now also populate GlueCEPolicyAssignedJobSlots for the GlueCE.
    • Note that torque server and client version have to be the same on a site for a proper interaction. Keep this in mind for the case of mixed SL4/SL5 torque client/server
    • Known Issue : problems in deployment of MPI together with torque_client. Fix currently in certification
  • Update to WMS
    • New version of ICE for the WMS to fix bug #59054
  • Update to CREAM
    • Bug fixes
    • With respect to the previous version this one fixes a couple of problems with the values published in the information service (in particular it enables monitoring to "fix installed capacity")
    • A fix for a vulnerability affecting BLAH is included as well
-- Early Adopter Sites for this update --
  • SiGNET (glite-CREAM_TORQUE, lcg-CE_TORQUE SL4, i386)
  • PPS-IFIC (glite-WMS, SL3, i386)
  • ITWM-PPS (glite-SE_dpm_mysql, WN_GLEXEC, SL4, x86_64)
  • WCSS-PPS (lcg-CE_TORQUE, SL44 x86_64
  • GUP-CERTIF-TB (lcg-CE_TORQUE, SL4, i386)
-- Reports -- -- Tentative date for release to production --
  • 8-Dec-09
-- Affected services currently NOT COVERED by the staged roll-out --
  • glite-MPI_utils (SL4, i386)
Please contact grid-deployment-managers@cern.ch to volunteer



PPS Coordination Report on 26-Nov-09

EGEE Pre-Production Service Coordination:


2009-11-25: Pilot of glexec/Argus in preparation
JRA1, SA3 and SA1 are organising a pilot service of the expected duration of two months focused on the new Site Authorization Service (Argus) in order to collect feedback from the experiments and to accelerate the testing and deployment in production of the new service.
A kick-off meeting with 5 interested sites was held today and an initial timeline was defined.
Sites interested in taking part in the activity are welcome to contact egee-pilot-argus@cern.ch .
Minutes of the kick-off : https://twiki.cern.ch/twiki/bin/view/LCG/PPIslandKickOff2009x11x25
Pilot Home page: https://twiki.cern.ch/twiki/bin/view/EGEE/PilotServiceArgus



PPS Coordination Report on 19-Nov-09

EGEE Pre-Production Service Coordination:


2009-11-18: release of Torque and CREAM in production
It was decided that the content of gLite 3.1 Update60 and gLite3.2 Update07, referenced below, will be released to production in the form of a staged roll-out. This is because two newly certified patched not previously released to PPS were added to the bundle. In addition to that the CREAM CE for SL5 is introduced plus an update of CREAM on SL4. So in practice the next bundle released to the EA sites will contain:
  • NEW Torque client and server for SL5 (64bit) (PATCHES:3097, 3074, 3052, 3404, 3481)
  • update to Torque client and server for SL4 (32 and 64bit) (PATCHES:2704, 2707, 2709, 2590, 3320, 3046, 3051)
    • Among others this version fixes a low-severity vulnerability
  • NEW CREAM for SL5 (PATCHES:3439, 3260, 3233)
  • update to CREAM for SL4 (PATCHES:3289, 3438)
    • With respect to the previous version this one fixes a couple of problems with the values published in the information service (in particular it enables monitoring to "fix installed capacity")
    • A fix for a vulnerability affecting BLAH is included as well
The update will be available in the preview pages (http://glite.web.cern.ch/glite/preview/default.asp) by the end of this week.

2009-11-17: staged roll-out of gLite security Update (Bundle 02)
Due to the issues found affecting the release for the UI and LB node types this update has been rejected and the early adopter sites rolled back. The fix will be proposed again in a later update (no date available)
"Preview" gLite release pages are available at: http://glite.web.cern.ch/glite/preview/default.asp
General info about staged roll-out available at: http://www.cern.ch/pps/index.php?dir=./rollout/

2009-10-16: Release of gLite 3.1 Update 60 to production in preparation
This update, to be expected within this week, will be released together with the SL5 version of Torque (gLite 3.2 Update07). It will contain:

  • Torque and Maui Patches introducing several bug fix plus fix for a vulnerability (PATCH:2590)(PATCH:2704)(PATCH:2707)(PATCH:2709)

2009-10-16: Release of gLite 3.2 Update 07 to production in preparation
This update, to be expected within this week, will be released together with a fix for the SL4 version of Torque (gLite 3.1 Update60).
It releases TORQUE server and TORQUE utils for the first time on SL5. Specifically:

2009-11-13: staged roll-out of gLite security Update (Bundle 02)
A new gLite Update was released to the Early Adopters sites (currently they are mostly PPS site) and it is now going through the due quarantine period. It introduces a security fix for a vulnerability affecting basically all grid node types.
Two issues have been found affecting the release for the UI and LB node types. The concerned rpms have been handed back to the developers and two sites (from Russia and SWE, thanks!) have been requested to roll-back to the previous version. Unfortunately, due to this incident the release date of the whole security update has slipped. The new date will be circulated as soon as available
"Preview" gLite release pages are available at: http://glite.web.cern.ch/glite/preview/default.asp
General info about staged roll-out available at: http://www.cern.ch/pps/index.php?dir=./rollout/



PPS Coordination Report on 12-Nov-09

EGEE Pre-Production Service Coordination:


2009-11-11: new gLite Update released as staged roll-out (Bundle 02)
A new gLite Update was released to the Early Adopters sites (currently they are mostly PPS site) and it is now going through the due quarantine period. It introduces a security fix for a vulnerability affecting basically all grid node types.
"Preview" gLite release pages are available at: http://glite.web.cern.ch/glite/preview/default.asp
General info about staged roll-out available at: http://www.cern.ch/pps/index.php?dir=./rollout/


EGEE-III PPS Quarterly Report Q6 for UE

Preproduction Service A slight re-organisation of the service took place in order to support the start-up of the staged roll-out process (see Middleware Deployment). All sites previously working in the deployment test have been requested to act as early adopters of the releases to production. One of the macroscopic effects of this re-oranisation is that since the middle of October the release to PPS done according to the traditional release process have been discontinued. The infrastructure is however ready to react should a release through PPS be required for whatever reasons. There have been no new pilot services requests in the past trimester and there are currently no active pilots. The plan for the next quarter is to start the final decommissioning of the PPS service by promoting a gradual re-absorbing of the PPS resources in the production infrastructure, possibly as fully qualified Early Adopters

Middleware Deployment There were 7 updates on gLite3.1(SL4) and two on gLite3.2 (SL5) for maintenance and bug fixes. Among the new features it's worth mentioning the back-porting to SL4 of the WMS 3.2 which among many other improvements notably has fixed the job submission path to the CREAM CE via WMS, previously broken. The roll-out of the new CREAM computing element in production is progressing, mainly driven by requirements coming from WLCG. at the endo of October the WLCG Management Board has finally recommended the sites to start flagging the new service as Production. The transition towards the new process for staged roll-out of middleware updates has started as planned with the initial participation of sites from PPS. The release procedure has been significantly changed and the first "staged" release happened at the end of October.The relevant documents are available at http://www.cern.ch/pps/index.php?dir=./rollout/ Finally, the middleware deployment and support team has been given the task of creating and maintaining a list of the recommended middleware versions for the production service, which is being compiled in collaboration with WOLCG. Plans for the next quarter include the final developments for the staged roll-out (integration with the GOCDB) and the consequent extension of the process on production scale; the extensive deployment of CREAM CEs at T2 sites and the completion of the recommended versions as well as the definition of the maintanance procedures for the list



PPS Coordination Report on 5-Nov-09

EGEE Pre-Production Service Coordination:


2009-11-05: gLite 3.1 Update 59 was released to production
The update introduces a new version of lcg-infosites enabling queries to multiple BDIIs
Release notes at: Test reports from staged roll-out at:

2009-10-28: First gLite Update released as staged roll-out
A first gLite Update was released to the Early Adopters sites (currently they are mostly PPS site) and it is now going through the due quarantine period. This is the first time the new procedure is used. Therefore initially we applied it to a very "soft" update, introducing l only a new version of the lcg-infosites in the production system.
"Preview" gLite release pages are available at: http://glite.web.cern.ch/glite/preview/default.asp
General info about staged roll-out available at: http://www.cern.ch/pps/index.php?dir=./rollout/
A question has been moved by some of the PPS sites currently involved in the exercise on how they should configure the nodes that currently point to PPS resources (e.g. PPS BDII, PPS WMS): for the first iterations of the new procedure these sites can leave these configuration parameter unchanged. There will be a warning when services will be put off-line. Alternatively nodes can be pointed to corresponding production services and so start the migration of the PPS nodes into the production system. This decision is fully supported from the PPS management but should however be shared by the PPS administrator with the production site and ROC management.



PPS Coordination Report on 29-Oct-09

EGEE Pre-Production Service Coordination:


2009-10-28: gLite 3.1 Update 58 was released to production
The content of the update is the same announced with Update 57 with exception of the BDII. Namely Release notes at:

2009-10-22: gLite 3.1 Update 57 was released to production
The update contained, among others, the GLUE 2.0 Enabled BDII (PATCH:3204;3208)
Incident in Production: gLite 3.1 update 57 caused SAM failures as outlined in GGUS ticket #52597. Approximately 30 CEs were not being published as the BDII service on those machines were in a stopped state. In addition, one site reported in GGUS ticket #52578 that their top-level BDII was missing entries from a site that we publishing information correctly.
On 23.10.2009 a broadcast was sent announcing the rollback, at the same time the gLite 3.1 release pages were updated with relevant information.
The reason for the BDII failures on the CE was that the conditional restart of the service failed. This was due to a problem with the rpm that was already installed and the problem was only triggered on the update. Most of the sites affected had auto-update enabled and hence updated the package automatically.
The problem with the top-level BDII was caused by the fact that the gLite 3.1 release was missing an rpm that was submitted for release on gLite 3.2 (patch 3154) but was never submitted for release on gLite 3.1. Release notes at:



PPS Coordination Report on 22-Oct-09

EGEE Pre-Production Service Coordination:


2009-10-21: gLite 3.1 Update 57 released to production
The update contains: Release notes at:

2009-10-19: New gLite Release Procedure in for staged roll-out
Within the transition towards the EGI operation model steps have been made to move from the current release workflow, based on the stage in PPS to a new model where updates are rolled-out directly in production in a controlled manner. This deployment model is commonly identified, as you have most likely heard, as ‘Staged Roll-out’. (http://www.cern.ch/pps/index.php?dir=./rollout/ )
In collaboration with SA3 we have developed a new release procedure to cope with the new model and now the plan is to start exercising it with those sites already involved in the release management as you are. This procedure fully replaces the PPS stage and it is meant to be applied in the future to the production sites that will have subscribed to the staged roll-out program.
The new procedure is referenced at https://twiki.cern.ch/twiki/bin/view/EGEE/ReleaseProcedure . Site and ROC representatives are invited to check it focusing especially on the steps relevant for ‘Early Adopter’ sites (marked in blue). Every comments will be welcome and appreciated.



PPS Coordination Report on 15-Oct-09

EGEE Pre-Production Service Coordination:


2009-10-16: Release of gLite 3.1 Update 57 in preparation
New versions of some middleware components are going to be released to the production in the week of the 19th October. Namely:

2009-10-12: gLite 3.2 Update06 was released to production
This update to PPS contains a new version of the VOBOX corresponding to a configuration tried out and validated by Alice off the standard release track Specifically:


  • 2009-10-06(1): Task: gL3.1 U53 --> PPS Details:kick-off e-mail sent (GGUS:52142) Status: In Progress . Antonio, Esteban

  • 2009-10-06(3): Task: gL3.1 U52 --> PPS Details: Release of Update 52 to PPS done and announced to the PPS site Admins (GGUS:51417) Status: Done . Antonio,Esteban

  • 2009-10-06(2): Task: gL3.1 U52 --> Requested the mirroring of the repository at CNAF (GGUS:51417) Status: In Progress . Antonio, Esteban


PPS Coordination Report on 8-Oct-09

EGEE Pre-Production Service Coordination:


2009-10-07: News from Cerification and Release teams:
  • Priority in the last three weeks was given to the preparation of a security fix for a vulnerability affecting an important number of node types in production.
  • The fixes are now certified and will be moved to PPS tomorrow 8/10
  • A corresponding update to production has to be expected next week
  • The application of the patch won't require a reconfiguration of the nodes but likely the services will need to be re-started

2009-09-06: Tests of Cream CE in production: in progress

  • CMS is testing the WMS-based submission to CREAM CES at CERN, CNAF, FZK and RAL
  • The tests so far haven't shown any performance limitation in ICE, the upgrade to the latest version, released with gLite 3.1 Update 55 is however recommended
  • A known issue in CREAM, reported with BUG:52651 "CREAM file descriptor overuse" was hit at CNAF
  • Supporting sites are now updating to the last version of CREAM (released on the 6th) that fixes the bug

2009-10-06: gLite 3.1 Update 56 released to production
This update conatains the CREAM CE version 1.5 (PATCH:3259)
Among others it fixes 2 high-risk vulnerabilities (1 potential root exploit) and brings many improvements to the robustness of CREAM services:

  • Vulnerability advisories:
  • VOMS server host certs are not required anymore in the CREAM CE node (the /etc/grid-security/vomsdir/*/*.lsc files are sufficient)
  • First release of new BLAH blparser for LSF and PBS/Torque (use of batch system status/history commands instead of parsing the log files)
  • IPv6 support for BLAH
  • Several bug fixes, in particular:
    • Fix for a problem with purge (not deleting the job sandbox dir in some cases)
    • Fix for a problem for which there could be at most 32000 active jobs per user
    • Fix for a problem for which the produced accounting logs (used by APEL and DGAS) was not properly filled if the CREAM CE is not the Torque server
Release notes: http://glite.web.cern.ch/glite/packages/R3.1/updates.asp

2009-10-06: gLite 3.1.0 PPS Update 52 went through deployment test in PPS
And it is now being installed on the remaining services
This update contains the CREAM CE version 1.5 (PATCH:3259)

  • VOMS server host certs are not required anymore in the CREAM CE node (the /etc/grid-security/vomsdir/*/*.lsc files are sufficient)
  • First release of new BLAH blparser for LSF and PBS/Torque (use of batch system status/history commands instead of parsing the log files)
  • IPv6 support for BLAH
  • Several bug fixes, in particular:
    • Fix for a problem with purge (not deleting the job sandbox dir in some cases)
    • Fix for a problem for which there could be at most 32000 active jobs per user
    • Fix for a problem for which the produced accounting logs (used by APEL and DGAS) was not properly filled if the CREAM CE is not the Torque server
  • Security fix

2009-10-05: SAM tests for CREAM available in the SAM Validation instance.

ROCs please check



PPS Coordination Report on 1-Oct-09

EGEE Pre-Production Service Coordination:


2009-09-30: Tests of Cream CE in production: in progress

  • CMS is testing the WMS-based submission to CREAM CES at CERN, CNAF, FZK and RAL
  • The tests so far haven't shown any performance limitation in ICE, the upgrade to the latest version, released with gLite 3.1 Update 55 is however recommended
  • A known issue in CREAM, reported with BUG:52651 "CREAM file descriptor overuse" was hit at CNAF

2009-09-22: gLite 3.2 Update 05 (SL5) went to production

  • Bugfixes in GFAL v1.11.8-2 and lcg_utilv1.7.6-1 (PATCH:3119)
  • new lcg-infosites version solving bug 38922 SL5/x86_64 (PATCH:3159)
  • SWAT (on WN) (PATCH:3172)
    • glite-swat (Site Wide Area Testing) is a replacement for grid-cm (Grid Configuration Monitoring)
  • GLUE 2.0Enabled BDII SL5/x86_64 (PATCH:3206)
  • YAIM core(PATCH:3106)
    • Several new variable added in site-info.def
  • YAIM Clients (PATCH:3106)
    • New site-info variables
    • New function to configure the new myproxy server and clients.
    • SL5 related bug fixes
This update will be released in preview on the 17th of September

2009-09-22: gLite 3.1 Update 55 went to production
This update contains:

  • Update of the ICE component of the WMS (PATCH:2862)
    • Improvements in performance and scalability
    • Improved use of memory
    • Use of jobdir instead of filelist as input for ICE
  • Bugfixes in GFAL v1.11.8-2 (PATCHES:3197,3198)
With respect to what previously announced the content of this update has been modified:
  • the release of CREAM 1.5 (PATCH:2666) has been postponed
  • Bugfixes to GFAL were added
This update will be released in preview on the 17th of September


  • 2009-09-18(1): Task: gL3.2 U08 --> PPS Details:kick-off e-mail sent (GGUS:51641) Status: In Progress . Esteban, Alvaro

  • 2009-09-18(3): Task: gL3.2 U07 --> PPS Details: Release of Update 07 to PPS done and announced to the PPS site Admins (GGUS:51434) Status: Done . Esteban, Alvaro

  • 2009-09-17(2): Task: gL3.2 U07 --> Requested the mirroring of the repository at CNAF (GGUS:51434) Status: In Progress . Esteban, Alvaro


PPS Coordination Report on 17-Sep-09

EGEE Pre-Production Service Coordination:


2009-09-17: release of gLite 3.2 Update 05 (SL5) to production in preparation

  • Bugfixes in GFAL v1.11.8-2 and lcg_utilv1.7.6-1 (PATCH:3119)
  • new lcg-infosites version solving bug 38922 SL5/x86_64 (PATCH:3159)
  • SWAT (on WN) (PATCH:3172)
    • glite-swat (Site Wide Area Testing) is a replacement for grid-cm (Grid Configuration Monitoring)
  • GLUE 2.0Enabled BDII SL5/x86_64 (PATCH:3206)
  • YAIM core(PATCH:3106)
    • Several new variable added in site-info.def
  • YAIM Clients (PATCH:3106)
    • New site-info variables
    • New function to configure the new myproxy server and clients.
    • SL5 related bug fixes
This update will be released in preview on the 17th of September

2009-09-16: release of gLite 3.1 Update 55 to production in preparation
This update will contain:

  • Update of the ICE component of the WMS (PATCH:2862)
    • Improvements in performance and scalability
    • Improved use of memory
    • Use of jobdir instead of filelist as input for ICE
  • Bugfixes in GFAL v1.11.8-2 (PATCHES:3197,3198)
With respect to what previously announced the content of this update has been modified:
  • the release of CREAM 1.5 (PATCH:2666) has been postponed
  • Bugfixes to GFAL were added
This update will be released in preview on the 17th of September

2009-09-15: gLite 3.1 baseline (SL4) : Staged roll-out
New versions of some middleware components are going to be released to the production sites in the next days with a controlled roll-out, Namely:

2009-09-14: Announce to the sites working in PPS deployment test and SA1 release testing:
Very soon the PPS release will be replaced by a "preview of the reelase to production" (in the preliminary discussion we are having with SA3 this should start already with the next scheduled production update for 3.1 - 56). So the next releases to production will contain both patches that have already gone through the deployment test and patches which are released to PPS for the first time. All the sites currently registered in deployment tests adn release testing are "de officio" registered for staged roll-out as well so they will receive new tasks to apply the update (unfortunately in this first transition time these tasks will include versions that some of them they have already tested in PPS but this will be a temporary thing). Sites will still have to provide reports as usual and the reports will be collected exactly as it happens now


  • 2009-09-08(1): Task: gL3.2 U07 --> PPS Details:kick-off e-mail sent (GGUS:51434) Status: In Progress . Esteban, Alvaro

  • 2009-09-08(3): Task: gL3.2 U06 --> PPS Details: Release of Update 06 to PPS done and announced to the PPS site Admins (GGUS:50594) Status: Done . Esteban, Alvaro

  • 2009-09-08(2): Task: gL3.2 U06 --> Requested the mirroring of the repository at CNAF (GGUS:50594) Status: In Progress . Esteban, Alvaro

  • 2009-09-08(2): Task: gL3.1 U52 --> PPS Details:kick-off e-mail sent (GGUS:51417) Status: In Progress . Esteban, Alvaro

  • 2009-09-08(3): Task: gL3.1 U51 --> PPS Details: Release of Update 51 to PPS done and announced to the PPS site Admins (GGUS:50907) Status: Done . Esteban, Alvaro

  • 2009-09-07(2): Task: gL3.1 U51 --> Requested the mirroring of the repository at CNAF (GGUS:50907) Status: In Progress . Esteban, Alvaro


PPS Coordination Report on 9-Sep-09

EGEE Pre-Production Service Coordination:


2009-09-08: gLite 3.2.0 PPS Update 07 was released to PPS
And it is now being tested by the deployment test team
This update to PPS contains a new version of the VOBOX corresponding to a configuration tried out and validated by Alice off the standard release track Specifically:

2009-09-08: gLite 3.2.0 PPS Update 06 went through PPS deployment test
This update to PPS contains a set of patches with updates to GFAL, yaim core and clients, BDII, lcg-infosites and Myproxy. It also releases TORQUE server and TORQUE utils for the first time in SL5. Specifically:

  • Torque client as server for SL5 (64bit) (PATCH:3074,3097)
  • R3.2 MyProxy 4.7 client compatible with myproxy server 4.7 (PATCH:3079)
  • gfal v1.11.8-2 and lcg_util v1.7.6-1 with bug fixes (PATCH:3119)
  • new lcg-infosites version solving bug 38922 SL5/x86_64 (PATCH:3159)
  • SWAT (on WN) (PATCH:3172)
    • glite-swat (Site Wide Area Testing) is a replacement for grid-cm (Grid Configuration Monitoring)
    • GLUE 2.0Enabled BDII SL5/x86_64 (PATCH:3206)
  • YAIM core(PATCH:3106)
    • Several new variable added in site-info.def
  • YAIM Clients (PATCH:3106)
    • feature to retrieve the subcluster-id from the tarball installations (PATCH:3110,3111)
Test reports available at http://www.cern.ch/pps/index.php?dir=./release/testreports/gLite3.2.0/gLite3.2.0-PPS-UPDATE06/&

2009-09-03: release of gLite 3.1 Update 55 to production in preparation
This update will contain:

  • CREAM CE 1.5 (PATCH:2666)
    • VOMS server host certs are not required anymore in the CREAM CE node (the /etc/grid-security/vomsdir/*/*.lsc files are sufficient)
    • First release of new BLAH blparser for LSF and PBBS/Torque (use of batch system status/history commands instead of parsing the log files)
    • IPv6 support for BLAH
    • Several bug fixes, in particular:
      • Fix for a problem with purge (not deleting the job sandbox dir in some cases)
      • Fix for a problem for which there could be at most 32000 active jobs per user
      • Fix for a problem for which the produced accounting logs (used by APEL and DGAS) was not properly filled if the CREAM CE is not the Torque server
  • Update of the ICE component of the WMS (PATCH:2862)
    • Improvements in performance and scalability
    • Improved use of memory
    • Use of jobdir instead of filelist as input for ICE



PPS Coordination Report on 3-Sep-09

EGEE Pre-Production Service Coordination:


2009-08-31: gLite 3.1 Update 54 was released to production
This update contains, among others:
  • Client nodes (UI, WN, VOBOX)
    • lcg-infosites: bug fix (PATCH:3158, 2605)
    • GFAL and lcg-utils: Bug fixes (PATCH:3043, 3048, 3115, 3120)
    • glexec - environment wrapper scripts (PATCH:3050)
  • Workload Management
    • LSF: glite-info-dynamic-scheduler-lsf introduces the parameter LSF_CE_PRODUCTION_STATUS (PATCH:2658)
    • APEL: bug fixes (PATCH:2755)
    • new globus-gma fixes an issue which under particular conditions may bring down the lcg-CE if overloaded (PATCH:3041)
    • [LCAS] lcas/lcas-interface update fixing globus related bug (PATCH:3084)
    • lcg-CE: cleanup-jobdirs cron job to remove stale jobdirs from LCG-CE (PATCH:3143)
    • YAIM CE: bug fix for correct interaction with Globus GMA (PATCH:3080)
  • Data Management
    • dCache 1.9.3 Client and dCap client: bug fixes (PATCH:3058, 3060)
  • YAIM: yaim core 4.0.8 : Bug fixes (PATCH:2971, 2981)

2009-08-28: gLite 3.1.0 PPS Update 51 was released to PPS
And it is now being tested by the deployment test team This update contains a new version of the BDII, FTS and CREAM CE + the ICE Plugin. In particular:

  • FTS2.2 for SL5 , client and server nodes (PATCH:2403, 2404)
    • SL5 ready
    • SRM-gridftp split
    • Logging to syslog according to Middleware Security Audit Logging Guidelines
    • Logging IP addresses in SRM calls
    • Removed obsolete commands: glite-transfer-placement-submit and
    • glite-transfer-discovery
    • Added site group management tools: glite-transfer-group-addmember
    • glite-transfer-group-list glite-transfer-group-removemember
    • Added channel management methods for new channel parameters
    • Merged transfer-api-c into transfer-cli.
    • DB maintenance packages are part of the core FTS package: history, purge
    • Compliant with the WLCG SRM MoU for spaces
  • CREAM CE 1.5 (PATCH:2666)
    • VOMS server host certs are not required anymore in the CREAM CE node (the /etc/grid-security/vomsdir/*/*.lsc files are sufficient)
    • First release of new BLAH blparser for LSF and PBBS/Torque (use of batch system status/history commands instead of parsing the log files)
    • IPv6 support for BLAH
    • Several bug fixes, in particular:
      • Fix for a problem with purge (not deleting the job sandbox dir in some cases)
      • Fix for a problem for which there could be at most 32000 active jobs per user
      • Fix for a problem for which the produced accounting logs (used by APEL and DGAS) was not properly filled if the CREAM CE is not the Torque server
  • Update of the ICE component of the WMS (PATCH:2862)
    • Improvements in performance and scalability
    • Improved use of memory
    • Use of jobdir instead of filelist as input for ICE
  • Change in glite -info-template (update of the 64 bit repo although implementing new functionality only for the lcg-CE) (PATCH:3036, 3037)
  • dCache security update dcache-server-1.9.1-11 (PATCH:3184, 3185)
    • Bug fixes + security update to prevent SQL injection issue in dCache
  • YAIM core
    • Several new variable added in site-info.def
  • YAIM Clients (PATCH:3104,3105)
    • feature to retrieve the subcluster-id from the tarball installations (PATCH:3110,3111)
  • SWAT (on WN) (PATCH:3168) * glite-swat (Site Wide Area Testing) is a replacement for grid-cm (Grid Configuration Monitoring)
  • GFAL bugfixes (PATCH:3197,3198)
  • GLUE 2.0 Enabled BDII (PATCH:3204)
Release notes: https://twiki.cern.ch/twiki/bin/view/EGEE/PPSReleaseNotes_310_PPS_Update51
Deployment test tasks sent and available at
http://www.cern.ch/pps/index.php?dir=./ActivityManagement/TaskTracker/


PPS Coordination Report on 27-Aug-09

EGEE Pre-Production Service Coordination:


2009-08-27: release of gLite 3.1 Update 54 to production in preparation
This update will contain, among others:
  • Client nodes (UI, WN, VOBOX)
    • lcg-infosites: bug fix (PATCH:3158, 2605)
    • GFAL and lcg-utils: Bug fixes (PATCH:3043, 3048, 3115, 3120)
    • glexec - environment wrapper scripts (PATCH:3050)
  • Workload Management
    • LSF: glite-info-dynamic-scheduler-lsf introduces the parameter LSF_CE_PRODUCTION_STATUS (PATCH:2658)
    • APEL: bug fixes (PATCH:2755)
    • new globus-gma fixes an issue which under particular conditions may bring down the lcg-CE if overloaded (PATCH:3041)
    • [LCAS] lcas/lcas-interface update fixing globus related bug (PATCH:3084)
    • lcg-CE: cleanup-jobdirs cron job to remove stale jobdirs from LCG-CE (PATCH:3143)
    • YAIM CE: bug fix for correct interaction with Globus GMA (PATCH:3080)
  • Data Management
    • dCache 1.9.3 Client and dCap client: bug fixes (PATCH:3058, 3060)
  • YAIM: yaim core 4.0.8 : Bug fixes (PATCH:2971, 2981)

2009-08-24: gLite 3.1 Update 53 was released to production
This update contains:

Release notes: http://glite.web.cern.ch/glite/packages/R3.1/updates.asp


PPS Coordination Report on 13-Aug-09

EGEE Pre-Production Service Coordination:


2009-08-12: release of gLite 3.1 Update 53 to production in preparation
This update will contain:

2009-07-24 gLite3.1.0 PPS Update 50 went to the deployment test and it is now being installed by the remaining PPS sites.
This update contains, among others:

  • host certificates for VOMS server voms.cern.ch
  • update to lcas/lcas-interface update to fix a globus related bug (PATCH:3084)
  • new version of lcg-infosites with a bug fix (PATCHES 3158, 2605 )
  • glite-info-dynamic-scheduler-lsf update package (PATCH:2658)
  • APEL and YAIM bug fixes (PATCH:2755)
  • globus-gma 1.0.13 (PATCH:3041)
  • GFAL 1.11.7 (PATCH:3043,2948)
  • gfal v1.11.8-1 and lcg_util v1.7.6-1 (PATCH:3115, 3120)
  • glexec - wrapper scripts to preserve pilot environment (PATCH:3050) glexec - wrapper scripts
  • dCache and dCap clients PATCHES:3058, 3060)
  • GLUE 2.0Enabled BDII (PATCH:3114, 3176) rejected
  • big fixes for WMS 3.2 (PATCH:3156)
Release notes: https://twiki.cern.ch/twiki/bin/view/EGEE/PPSReleaseNotes_310_PPS_Update50 Test reports: http://egee-pre-production-service.web.cern.ch/egee-pre-production-service/index.php?dir=./release/testreports/gLite3.1.0/&


  • 2009-07-28(1): Task: gL3.2 U06 --> PPS Details:kick-off e-mail sent (GGUS:50594) Status: In Progress . Esteban, Alvaro

  • 2009-07-28(3): Task: gL3.2 U05 --> PPS Details: Release of Update 05 to PPS done and announced to the PPS site Admins (GGUS:50397) Status: Done . Esteban, Alvaro

  • 2009-07-28(2): Task: gL3.2 U05 --> Requested the mirroring of the repository at CNAF (GGUS:50397) Status: In Progress . Esteban, Alvaro


PPS Coordination Report on 30-Jul-09

EGEE Pre-Production Service Coordination:


2009-07-29: gLite 3.1 Update 52 was released production
This update contains:
  • 5th update to CREAM-CE (PATCH:3042) (PATCH:2782)
    • Fixed a problem concerning publication in the BDII of FQAN voviews and DENY tags
    • CREAM and CEMon services publish themselves in the Information Service
    • Support for new blparser for LSF and PBS/Torque (not released yet) implemented. Since the new parser has not been released yet, don't try it
    • Implemented a workaround for the LB-locallogger restart problem
  • Addition of python-ldap as dependency to WN (PATCH:3024)(PATCH:3023)
  • new host certificates for the VOMS server voms.cern.ch (PATCH:2933) (PATCH:3094)

2009-07-27: gLite 3.2 Update 04 was released to production
This contains

  • BDII (PATCH:2959) + fix for BUG:53521
    • This version fixes an issue observed with glite-BDII 5.0.0-22 at some production sites and reported with GGUS:50372: it was noticed that in particular conditions site could disappear from the top-level BDII if the 'GlueTop' attribute is removed from the services they publish.
      If you have this version deployed at your top-level BDII, please upgrade asap
  • LFC mysql(PATCH:2924) for SL5
  • DPM mysql(PATCH:3020) for SL5
  • GFAL libraries(PATCH:3049) for SL5.
  • new versions of yaim core (PATCH:2982)and yaim clients (PATCH:2962).
    • groups.d structure is now available for all the node types.
    • various bug fixes
  • WN

2009-07-24 gLite3.2 PPS Update 05 was released to PPS and it is going through the deployment test .
This update contains:

  • PATCH:4154 with a fix for the BDII for an issue that has been detected at some production sites (GGUS:50372).
The fix is to be applied to the top level BDIIs.



EGEE-III PPS Quarterly Report Q5 for UE

Pre-Production Service

Middleware Pilot Services: Preproduction Service Pilot Services : SCAS/glexec : the main issue found by sites and VOs during the pilot were followed-up in collaboration with the WLCG GDB . As a result the functionality was released to production at the beginning of June. This version includes the two main features requested during the pilot phase: a more detailed set of error codes and a method to transparently handle the user environment. The pilot is virtually closed. The deployment issues experienced at IN2P3 are being followed up by the developers in a separate effort. CREAM/ICE : stress-testing activity started in collaboration with SA3 in order to verify some of the transition criteria defined by the GDB. Interesting findings produced in June were fed back to developers. Two new piloting activities started: WMS 3.2 at CERN and a WLCG package to address special requirements of the HEP experiments on SL5 (FZK, OXFORD, GRIF). Deployment test: at the beginning of July the ROCs contributed resources to double the layout of the deployment test infrastructure in order to support the upcoming services on SLC5. For the next quarter a slight re-organisation of the service is foreseen in support to the start-up of the staged roll-out process (see Middleware Deployment)

Middleware Deployment There were 9 updates on gLite3.1(SL4) for maintenance and bug fixes. Important new features released were the plugin for the WMS to submit to ARC-CE, the Site Central Authorization Service (SCAS) and a new version of glexec. On gLite3.2 (SL5) UI, BDII DPM and LFC for SL5 were released with 3 updates. A first version of the WN for Debian was released in June. In view of the transition to a new distribution model within EGI an extensive staged roll out of middleware updates was deemed necessary by both release managers and site/regional operators in order to protect the production environment against possibly broken middleware components. An intense activity took place during this quarter both within SA1 and SA3 in order to

  1. produce requirements and guidelines (SA1)
  2. draft an high level definition of the process (SA1, SA3)
  3. define workplan and timelines for the transition of the infrastructure to the new model (SA1)
The relevant documents are available at http://www.cern.ch/pps/index.php?dir=./rollout/ During the next quarter a significant part of the Middleware Deployment and Support activity will be devolved to the implementation of this process which, in the initial phases, will be instrumented essentially via a re-conversion of the PPS resources currently supporting the “deployment test” and “release testing” activities.



PPS Coordination Report on 23-Jul-09

EGEE Pre-Production Service Coordination:


2009-07-21: Release of gLite 3.2 Update 04 to production in preparation
The update will be available at the beginning of next week and will include This update will contain

2009-07-20 gLite3.2 PPS Update 04 passed successfully the deployment test in PPS
This update contains new releases for BDII (PATCH:2959), LFC mysql(PATCH:2924) and DPM mysql(PATCH:3020) and GFAL libraries(PATCH:3049) in SL5. It also contains new versions of yaim core (PATCH:2982)and yaim clients(PATCH:2962). Furthermore now python-ldap is added as a dependency to WN (PATCH:3023)
Test reports available at http://egee-pre-production-service.web.cern.ch/egee-pre-production-service/index.php?dir=./release/testreports/gLite3.2.0/gLite3.2.0-PPS-UPDATE04/

2009-07-15: gLite 3.1 Update 51 was released to production
This update contains

  • New release of yaim clients with
    • several bug fixes
    • New site-info.def variables
  • New version of DPM (1.7.2-4)
  • New version of LFC (1.7.2-4) + YAIM


  • 2009-07-20(1): Task: gL3.2 U05--> PPS Details:Ticket created on ggus(GGUS:50397) and kick-off e-mail sent to gLite Release Team Status: In Progress . Alvaro, Esteban

  • 2009-07-20(3): Task: gL3.2 U04 --> PPS Details:Bulletin created, broadcast sent to the PPS site admins ggus(GGUS:49412) Status: Done . Alvaro,Esteban

  • 2009-07-15(2): Task: gL3.2 U04--> PPS Details:Mirror requested at CNAF ggus(GGUS:49412) Status: In Progress . Alvaro,Esteban

  • 2009-07-07(1): Task: gL3.1 U50--> PPS Details:Ticket created on ggus(GGUS:50069) and kick-off e-mail sent to gLite Release Team Status: In Progress . Alvaro, Esteban

  • 2009-07-07(3): Task: gL3.1 U49 --> PPS Details:Bulletin created, broadcast sent to the PPS site admins ggus(GGUS:49413) Status: Done . Alvaro,Esteban

  • 2009-07-06(2): Task: gL3.1 U49--> PPS Details:Mirror requested at CNAF ggus(GGUS:49413) Status: In Progress . Alvaro,Esteban


PPS Coordination Report on 29-Jun-09

EGEE Pre-Production Service Coordination:


2009-06-29: Pilot service of glexec/SCAS: in progress
  • A new version of lcas (glite-security-lcas-1.3.11-2) was released to the pilot and it is now being installed by the three participant sites. It fixes BUG:52417 in lcas, which led to a segfault when the proxy file was corrupt (issue seen by Atlas at LANCASTER). This is only a change is lcas, not lcas-interface

2009-06-29: Release of gLite 3.1 Update 48 to production has been requested
The exact release date is not available yet but the update should be out within the next two weeks. It will contain:

  • First release of Site Central Authorization Service (SCAS) + glexec + LCAS/LCMAPS
  • LFC version 1.7.2-3 (PATCHES:2969,2970,2988,2992)
    • sources, configuration and init scripts have been improved for portability to SL5, Debian, OS X and Solaris.
    • Two new methods have been added to the LFC server and the API for the bulk removal or addition of entries.
  • YAIM for lcg-CE: glite-yaim-lcg-ce 4.0.5 supporting the configuration for the gatekeeper to publish info about installed capacity (publish the gatekeeper and the RTEpublisher services) (PATCHES:2840, 2841,2922)



PPS Coordination Report on 18-Jun-09

EGEE Pre-Production Service Coordination:


2009-06-25: gLite 3.1 Update 48 was released to production
The update contains: As usual a preview release page will be made available to the sites participating to the release testing one day before the scheduled date.



PPS Coordination Report on 18-Jun-09

EGEE Pre-Production Service Coordination:


2009-06-17: Release of gLite 3.1 Update 48 to production has been requested
The exact release date is not available yet but the update should be out within the next two weeks. It will contain: As usual a preview release page will be made available to the sites participating to the release testing one day before the scheduled date.

2009-06-17: Important : Change in the gLite release procedure
In accordance to the software management process as it is envisaged within the EGI model, in the following months the deployment test at the PPS sites will gradually be replaced by the "early adoption" process (aka "staged roll-out") . Therefore already with the next releases of gLite some patches will bypass the PPS, meaning that they won't be object of deployment test anymore.
This approach will be used for Debian WN, BDII and LFC on the baseline gLite 3.2 (SL5 OS).
The PPS will continue providing support (namely deployment test) for all the updates to gLite 3.1 (based on SL4)

2009-06-16: gLite 3.2 Update 02 was released to production
The update contains

  • first version of the gLite UI for SLC5 (PATCH:3035)

2009-06-16: Pilot service of Cream CE: in progress



PPS Coordination Report on 11-Jun-09

EGEE Pre-Production Service Coordination:


2009-06-09: Release of gLite 3.1 Update 47 to production has been requested
The update scheduled for the 15th of June, will contain:

  • Resource BDII in glite-AMGA + improvements in YAIM for AMGA (PATCH:2646)
  • MyProxy client on WN (requested by Atlas for PanDA) (PATCH:2838)(PATCH:2849)
  • lcg_util 1.7.4 and GFAL1.11.6 fixing the BUG:50265 (core dump). PATCHES:3014/3015/3029/3031
    • Unfortunately this version suffer from the regression of an old bug, but this is however seen as an improvement with respect to the current conditions of the production system.
Two features previously announced have been postponed, namely As usual a preview release page will be made available to the sites participating to the release testing one day before the scheduled date.

2009-06-09: gLite 3.2 PPS Update 03 went through deployment testing and it is now being propagated to the whole PPS
The update contains

  • first version of the gLite UI for SLC5 (PATCH:3035)

2009-06-09: gLite 3.1 PPS Update 48 went through deployment testing and it is now being propagated to the whole PPS
The update contains

  • New version of GFAL (1.11.6) and lcg_util (1.7.4) notably fixing BUG:50265 (segmentation fault) (PATCHEs: 3031,3029,3015,3014)
  • New version of glite-info-provider-service for the lcg CE (PATCH:2841) to publish the gatekeeper and the RTEpublisher services.
  • Updated globus-gma (PATCH:2749)
    • fixes GGUS ticket 42981 (defunct globus-gma processes, from LAL)
    • workaround for a Condor-G bug (bug #40359),(Condor-G does not distinguish between different VOMS groups, roles or VOs when the proxies have the same DN)


  • 2009-06-10(1): Task: gL3.1 U49--> PPS Details:Ticket created on ggus(GGUS:49413) and kick-off e-mail sent to gLite Release Team Status: In Progress . Alvaro, Esteban

  • 2009-06-10(1): Task: gL3.2 U04--> PPS Details:Ticket created on ggus(GGUS:49412) and kick-off e-mail sent to gLite Release Team Status: In Progress . Alvaro, Esteban

  • 2009-06-10(3): Task: gL3.1 U48 --> PPS Details:Bulletin created, broadcast sent to the PPS site admins ggus(GGUS:49164) Status: Done . Alvaro,Esteban

  • 2009-06-10(3): Task: gL3.2 U03- --> PPS Details:Bulletin created, broadcast sent to the PPS site admins ggus(GGUS:49252) Status: Done . Alvaro,Esteban

  • 2009-06-09(2): Task: gL3.2 U03--> PPS Details:Mirror requested at CNAF ggus(GGUS:49252) Status: In Progress . Alvaro,Esteban

  • 2009-06-09(2): Task: gL3.1 U48--> PPS Details:Mirror requested at CNAF ggus(GGUS:49164) Status: In Progress . Alvaro,Esteban


PPS Coordination Report on 11-Jun-09

EGEE Pre-Production Service Coordination:


2009-06-09: Release of gLite 3.1 Update 47 to production in preparation
The update, will contain:

In addition to that we will release a version of lcg_util anf GFAL fixing the BUG:50265 (core dump). Unfortunately this version suffers from the regression of an old bug, but this is however seen as an improvement with respect to the current conditions of the production system.
Two features previously announced have been postponed, namely An exact schedule for this release is not available yet. It should happen within this week, but it will wait for the new version of GFAL and lcg_util to be carefully checked in the PPS deployment test.
As usual a preview release page will be made available to the sites participating to the release testing one day before the scheduled date.

2009-06-05: gLite 3.2 PPS Update 03 was released to PPS and it is currently in phase of deployment testing
The update contains

  • first version of the gLite UI for SLC5 (PATCH:3035)

2009-06-04: gLite 3.1 PPS Update 48 was released to PPS and it is currently in phase of deployment testing
The update contains

  • New version of GFAL (1.11.6) and lcg_util (1.7.4) notably fixing BUG:50265 (segmentation fault) (PATCHEs: 3031,3029,3015,3014)
  • New version of glite-info-provider-service for the lcg CE (PATCH:2841) to publish the gatekeeper and the RTEpublisher services.
  • Updated globus-gma (PATCH:2749)
    • fixes GGUS ticket 42981 (defunct globus-gma processes, from LAL)
    • workaround for a Condor-G bug (bug #40359),(Condor-G does not distinguish between different VOMS groups, roles or VOs when the proxies have the same DN)


  • 2009-06-04(1): Task: gL3.2 U03--> PPS Details:Ticket created on ggus(GGUS:49252) and kick-off e-mail sent to gLite Release Team Status: In Progress . Alvaro, Esteban

  • 2009-06-04(3): Task: gL3.2 U02 --> PPS Details:Bulletin created, broadcast sent to the PPS site admins ggus(GGUS:47326) Status: Done . Alvaro,Esteban


PPS Coordination Report on 4-Jun-09

EGEE Pre-Production Service Coordination:


2009-06-02: Release of gLite 3.1 Update 47 to production has been requested
The update, will contain:

  • Resource BDII in glite-AMGA + improvements in YAIM for AMGA (PATCH:2646)
  • Improvements to fetch-crl script (PATCH:2778) (PATCH:2780)
  • MyProxy client on WN (requested by Atlas for PanDA) (PATCH:2838)(PATCH:2849)
  • WN: grid-cm-* packages provide worker node configuration monitoring published on the Active MQ messaging system (PATCH:2660 PATCH:2661)
    • Note that the job monitoring on the WN is enabled by default. The publisher is designed to be very lightweight and fault tolerant and it doesn't store data locally. However instructions to disable it will be provided in the release notes.
A schedule for this release is not available yet. In fact the release preparation team is currently busy with other priorities, namely
  • Debian 4 WN
  • SL5 UI release to PPS
  • glite 3.1 release to PPS (lcg_util/GFAL)
As usual a preview release page will be made available to the sites participating to the release test one day before the scheduled date.

2009-06-02: gLite 3.2 PPS Update 02 went through deployment test and it is now being installed by the remaining PPS sites
The update contains

2009-05-29: gLite 3.1 PPS Update 47 went through the deployment test and it is now being installed by the remaining PPS sites
The update contains

  • LFC v1.7.2-4 + YAIM (PATCH:2969)(PATCH:2970)(PATCH:2992)(PATCH:2988)
    • Two new methods have been added to the LFC server and the API for the bulk removal or addition of entries.
  • Cream and CEMon clients on VOBOX (PATCH:2996)
  • submission to ARC CE enabled on WMS (PATCH:2964)
  • WMS 3.2 for SLC4 + YAIM (PATCH:2954)(PATCH:2597)(PATCH:2896)
    • important code re-structuring featuring:
      • parallel match-making
      • ISM: restructured algorithm for matchmaking in case of data requirements specified in the JDL.
      • added support for IPv6
      • improved error reporting for DAGs
      • run-time selection of LB type: server or proxy.
      • the jobwrapper template is now cached at each WM start (restart to re-load any change)
      • restructured jobwrapper (also removed perl dependencies)
      • dumping the ISM can be done more often at a lesser cost, simply by creating a jobdir request (basically a file) like this one: [ command = "ism_dump"; ]
      • this code baseline is ready to support Grid Site Delegtion 2, at the moment disabled given that for backward-compatibility needs coming from external packages the project has been build agains Gridsite 1.1.18
    • Working version of ICE for submission to CREAM
    • Release notes in https://twiki.cnaf.infn.it/cgi-bin/twiki/view/EgeeJra1It/ReleaseNotes2597
  • dCache 1.9.1-7 Sever and 1.9.0-9 Client release (PATCH:2868)(PATCH:2869)
    • Info service and info provider
    • Permanent registration of deleted files
    • SRM Client Tools
    • Log4j Runtime User Interface
    • dCacheConfigure Script
  • PATCH:2922 New version of glite-info-templates introducing the Glue attribute GlueHostProcessorOtherDescription
  • PATCH:2841 New version of glite-info-provider-service to publish the gatekeeper and the RTEpublisher services.
  • PATCH:2851 several minor fixes for lcg-info-dynamic-scheduler (REJECTED)
  • glite-yaim-lcg-ce 4.0.5 featuring the changes in the infosys on the CE + other bug fixes (PATCH:2840)
  • Torque and Maui Patches fixing a vulnerability (PATCH:2590)(PATCH:2704)(PATCH:2707)(PATCH:2709)


Older Posts ...

Edit | Attach | Watch | Print version | History: r610 < r609 < r608 < r607 < r606 | Backlinks | Raw View | WYSIWYG | More topic actions
Topic revision: r610 - 2010-03-25 - unknown
 
    • Cern Search Icon Cern Search
    • TWiki Search Icon TWiki Search
    • Google Search Icon Google Search

    LCG All webs login

This site is powered by the TWiki collaboration platform Powered by PerlCopyright &© 2008-2019 by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
Ideas, requests, problems regarding TWiki? Send feedback