Alert.png The wiki is deprecated and due to be decommissioned by the end of September 2022.
The content is being migrated to other supports, new updates will be ignored and lost.
If needed you can get in touch with EGI SDIS team using operations @ egi.eu.

EGI QualityCritera Release 4 Comments

From EGIWiki
Jump to navigation Jump to search

Comments to QC4

IGE

GridWay, Reviewers: Eduardo Huedo, Ismael Marin Carrion, UCM

AUTHN_CRED_1 - fine

AUTHN_PROTO_1 - fine but just used to interact with the middleware

AUTHN_DELEG_1 - fine


PARALLEL_MPI_1 – fine

PARALLEL_MPI_2 – fine

PARALLEL_OMP_1 – fine

PARALLEL_OMP_2 – fine


JOBSCH_IFACE_1 - fine

JOBSCH_EXEC_1 - fine

JOBSCH_EXEC_2 - fine

JOBSCH_JOB_1 - fine

JOBSCH_JOB_2 - fine

JOBSCH_JOB_3 - fine

JOBSCH_JOB_4 – fine (non-mandatory)

JOBSCH_JOB_5 - fine

JOBSCH_JOB_6 – fine (non-mandatory)

JOBSCH_JOB_7 – fine (non-mandatory) but not supported

JOBSCH_JOB_8 – fine (non-mandatory)

JOBSCH_SERVICE_1 – fine (non-mandatory)

JOBSCH_SERVICE_2 – N/A (non-mandatory)

JOBSCH_SERVICE_3 – fine (non-mandatory)

JOBSCH_SERVICE_4 – fine (non-mandatory)

JOBSCH_SERVICE_5 – fine (non-mandatory)


MON_PROBE_JOBSCH_1 – N/A (WMS only) but could be done, if needed,

through gsissh probes


OGSA-DAI, Reviewer: Charaka Palansuriya, EPCC

QC Criteria Documents available at: https://documents.egi.eu/public/ShowDocument?docid=1153

DATAACCESS_API_1

- This is *not* applicable to OGSA-DAI REST.  Note that there is no functionality relationship between WS-DAI and OGSA-DAI

DATAACCESS_OGSADAI_1

- This is subject to the following.
      * Data resources with SQL are supported
      * XML and files are not explicitly supported any more due to lack of interests from users
      * Deploy remote resource is *not* available
      * Undeploy resource is not explicitly supported
      

DATAACCESS_OGSADAI_2

- This will not be available in the OGSA-DAI REST 

DATAACCESS_OGSADAI_3

- This is fine

DATAACCESS_OGSADAI_4

- This is fine


General Review of all QC

Reviewers: Steve Crouch, Tim Parkinson, John Robinson, SOTON


GENERIC_DOC_1 (non-mandatory) GENERIC_DOC_2 - fine GENERIC_DOC_3 (non-mandatory) GENERIC_DOC_4 (non-mandatory) GENERIC_DOC_5 (non-mandatory) GENERIC_DOC_6 (non-mandatory) GENERIC_DOC_7 (non-mandatory) GENERIC_DOC_8 - fine GENERIC_DOC_9 - Why is this documentation? It's true, but for actual testing? Is this the same as GENERIC_SOFT_2?

GENERIC_DOC_10 - fine GENERIC_DOC_11 - what is explicitly meant by policy? How is this related to IGE?

GENERIC_DIST_1 (non-mandatory) GENERIC_DIST_2 - where is it?? GENERIC_DIST_3 - is only SL5 of interest to UMD or also other distributions? GENERIC_SOFT_1 - fine GENERIC_SOFT_2 - fine

GENERIC_SERVICE_1 - fine GENERIC_SERVICE_2 - fine GENERIC_SERVICE_3 (non-mandatory) GENERIC_SERVICE_4 (non-mandatory) GENERIC_SERVICE_5 (non-mandatory) GENERIC_SERVICE_6 - fine


GENERIC_SEC_1 - Does it really mean all files/dirs or just logs/config, etc. files? GENERIC_SEC_2 - fine GENERIC_SEC_3 - fine

GENERIC_MISC_1 - where is it?? GENERIC_MISC_2 – fine

AUTHN_CRED_1 - fine AUTHN_PROTO_1 - For EGI: Makes sense, but SSLv3? Not commonly used - For EGI: Inconsistent pass/fail criteria - no mention of SSLv3 for pass

AUTHN_DELEG_1 - Need to clarify. AUTHN_CA_1 – how is this related to IGE? AUTHN_CA_2 – how is this related to IGE?

AUTHN_CA_3 – how is this related to IGE? ATTAUTH_IFACE_1 - fine ATTAUTH_IFACE_2 - fine ATTAUTH_IFACE_3 - fine ATTAUTH_IFACE_4 (non-mandatory)

ATTAUTH_MGMT_1 - fine ATTAUTH_MGMT_2 - fine ATTAUTH_MGMT_3 - fine, 'Attributed' should be 'attributes' ATTAUTH_MGMT_4 - fine ATTAUTH_MGMT_5 – fine

ATTAUTH_MGMT_6 - fine

ATTAUTH_WEB_1 – how is this related to IGE? ATTAUTH_WEB_2 - how is this related to IGE? ATTAUTH_WEB_3 - how is this related to IGE? ATTAUTH_WEB_4 - how is this related to IGE? ATTAUTH_WEB_5 - how is this related to IGE?


AUTHZ_MGMT_1 – how is this related to IGE? AUTHZ_MGMT_2 - how is this related to IGE? AUTHZ_PCYDEF_1 - fine AUTHZ_PCYDEF_2 - fine- 'XAMCL' should be 'XACML' AUTHZ_PCYDEF_3 (non-mandatory) AUTHZ_PCYDEF_4 – fine

AUTHZ_PEP_2 – fine (needs to be ckecked how it is done with IGE software)


CREDMGMT_IFACE_1 - fine, but - What about proxies explicitly? CREDMGMT_IFACE_2 - fine CREDMGMT_IFACE_3 - fine

CREDMGMT_LINK_1 (non-mandatory)

JOBEXEC_IFACE_1 - fine

JOBEXEC_JOB_1 - fine JOBEXEC_JOB_2 - fine JOBEXEC_JOB_3 - fine

JOBEXEC_EXECMNGR_1 - fine JOBEXEC_EXECMNGR_2 – fine

JOBEXEC_EXECMNGR_3 - fine

JOBEXEC_AVAIL_1 - For EGI: Not clear on the criteria...? Does this mean that it is required that more than one gatekeepers to be able to submit to the same batch system?

JOBEXEC_AVAIL_2 (non-mandatory) JOBEXEC_AVAIL_3 - where is it?? JOBEXEC_AVAIL_4 (non-mandatory)

PARALLEL_JOB_1 - fine PARALLEL_JOB_2 (non-mandatory)

PARALLEL_JOB_3 (non-mandatory) – This is not supported on all the EGI-accepted batch systems. It's also dependent on the batch-system configuration.

PARALLEL_MPI_1 - fine PARALLEL_MPI_2 – fine (with additional scripts) PARALLEL_OMP_1 - fine PARALLEL_OMP_2 – fine (with additional scripts) INTERACTIVE_JOB_1 (non-mandatory) INTERACTIVE_JOB_2 (non-mandatory) INTERACTIVE_JOB_3 (non-mandatory) INTERACTIVE_JOB_4 (non-mandatory)

JOBSCH_IFACE_1 - fine

JOBSCH_EXEC_1 – fine - 'an' should be 'a' JOBSCH_EXEC_2 – fine JOBSCH_JOB_1 - fine JOBSCH_JOB_2 - fine JOBSCH_JOB_3 - fine JOBSCH_JOB_4 (non-mandatory) JOBSCH_JOB_5 – fine

JOBSCH_JOB_6 (non-mandatory) JOBSCH_JOB_7 (non-mandatory) JOBSCH_JOB_8 (non-mandatory)

JOBSCH_WMS_1 (N/A) JOBSCH_WMS_2 (N/A) JOBSCH_WMS_3 (N/A) JOBSCH_WMS_4 (N/A)

JOBSCH_WMS_SEC_1 (N/A)

JOBSCH_WMS_BUG_1 (N/A) JOBSCH_WMS_BUG_2 (N/A)

JOBSCH_SERVICE_1 (non-mandatory) JOBSCH_SERVICE_2 (non-mandatory) JOBSCH_SERVICE_3 (non-mandatory) JOBSCH_SERVICE_4 (non-mandatory) JOBSCH_SERVICE_5 (non-mandatory)

INFOMODEL_SCHEMA_1 - fine, but IIS - supports GLUE2.0

INFOMODEL_SCHEMA_2 (non-mandatory)

INFO_DISC_IFACE_1 – don’t know

INFODISC_AGG_1 (non-mandatory) INFODISC_AGG_2 - fine INFODISC_AGG_3 – fine - This requirement MUST be concrete. It's a mandatory requirement without specific dynamic information stated. Information that can be published is specified by the GLUE schema in form of mandatory or optional class attributes. If EGI wants to extend the list of mandatory attributes this must be described in detail in the requirement.

INFODISC_AVAIL_1 – fine - Since this is mandatory, some more detail would be appreciated. For instance specifying the data size that should be handled.

MSG_IFACE_1 – fine

METADATA_LFC_API_1 - N/A, specific to gLite LFC METADATA_LFC_FUNC_1 - N/A, specific to gLite LFC METADATA_LFC_FUNC_2 - N/A, specific to gLite LFC METADATA_LFC_FUNC_3 - N/A, specific to gLite LFC METADATA_LFC_FUNC_4 - N/A, specific to gLite LFC METADATA_LFC_FUNC_5 - N/A, specific to gLite LFC

METADATA_AMGA_API_1 - N/A, specific to AMGA METADATA_AMGA_API_2 - N/A, specific to AMGA

METADATA_AMGA_FUNC_1 - N/A, specific to AMGA METADATA_AMGA_FUNC_2 - N/A, specific to AMGA METADATA_AMGA_FUNC_3 - N/A, specific to AMGA METADATA_AMGA_FUNC_4 - N/A, specific to AMGA METADATA_AMGA_FUNC_5 - N/A, specific to AMGA

FILECRYPT_KEY_1 - N/A, specific to gLite Hydra FILECRYPT_KEY_2 - N/A, specific to gLite Hydra FILECRYPT_KEY_3 - N/A, specific to gLite Hydra

FILECRYPT_FILE_1 - N/A, specific to gLite Hydra FILECRYPT_FILE_2 - N/A, specific to gLite Hydra

FILEACC_API_1 (non-mandatory) FILEACC_API_2 (non-mandatory)

FILETRANS_API_1 - fine FILETRANS_API_2 - fine FILETRANS_API_3 – fine

FILETRANSFSCH_CHANNEL_1 - N/A, gLite FTS specific FILETRANSFSCH_CHANNEL_2 - N/A, gLite FTS specific FILETRANSFSCH_CHANNEL_3 - N/A, gLite FTS specific

FILETRANSFSCH_MGMT_1 - N/A, gLite FTS specific FILETRANSFSCH_MGMT_2 - N/A, gLite FTS specific - Spelling mistake 'file transfer operationS' FILETRANSFSCH_MGMT_3 - N/A, gLite FTS specific

STORAGE_API_1 - fine STORAGE_API_2 – fine - Poorly worded criteria in description; it's specified as a procedure

STORAGE_DEVICE_1 STORAGE_DEVICE_2 (non-mandatory) STORAGE_DEVICE_3 (non-mandatory) –

STORAGE_DEVICE_4 (non-mandatory) - Description says 'must support' but it's not mandatory!

INSTRUMENT_IE_1 - N/A INSTRUMENT_IE_2 - N/A INSTRUMENT_IE_3 - N/A INSTRUMENT_IE_4 - N/A

MON_NCG_1 – fine (if it refers to Globus Nagios probles) MON_NCG_2 – fine (if it refers to Globus Nagios probles)


MON_PORTAL_1 - N/A MON_PORTAL_2 - N/A MON_PORTAL_3 - N/A MON_PORTAL_4 - N/A MON_PORTAL_5 - N/A MON_PORTAL_6 - N/A

MON_DB_1 - N/A MON_DB_2 - N/A MON_DB_3 - N/A

MON_PROBE_1 MON_PROBE_2

Are these probes are needed for components that are registered in GOCDB?

MON_PROBE_JOBEXEC_1

MON_PROBE_JOBEXEC_2 - N/A, CREAM only MON_PROBE_JOBEXEC_3 – N/A - This should not be mandatory as this requirement seems to be specific to the gLite infrastructure and VERY intrusive for the resource provider.

MON_PROBE_JOBSCH_1 - N/A, WMS only

MON_PROBE_STORAGE_1 - N/A, SRM only

MON_PROBE_METADATA_1 - N/A, LFC only

ACC_JOBEXEC_1 – fine - 'Support is expected for...' - this is a mandatory criterion, but with a non-mandatory set of LRMs to support, i.e. must support: Torque, SGE, Condor, LSF

ACC_STORE_1 - Input from TP section states records must include following checklist of items (e.g. LCG data), but what about non-LCG jobs?

ACC_STORE_2 – GridSAFE?

ACC_CRON_1 - Local in this context means site? Inconsistent with previous terminology

ACC_CRON_2 - Global in this context means central? Inconsistent with previous terminology

ACC_PORTAL_1 - fine ACC_PORTAL_2 – fine

ACC_PORTAL_3 - fine

ACC_PORTAL_4 - fine ACC_PORTAL_5 - fine ACC_PORTAL_6 – fine

CLIENT_TOOLS_1 (non-mandatory) CLIENT_TOOLS_2 (non-mandatory)

CLIENT_API_1 - N/A, applies to SAGA only CLIENT_API_2 (non-mandatory)

CLIENT_API_BES_1 - N/A, applies to SAGA only

VIRT_MGMT_API_1 - fine

VIRT_MGMT_OPS_1 - fine VIRT_MGMT_OPS_2 - fine VIRT_MGMT_OPS_3 - fine VIRT_MGMT_OPS_4 - fine

VIRT_IMG_1 (non-mandatory)

VIRT_IMGDIST_1 - N/A, for StratusLab VIRT_IMGDIST_2 - N/A, for StratusLab VIRT_IMGDIST_3 - N/A, for StratusLab VIRT_IMGDIST_4 - N/A, for StratusLab

Other Notes:

Sections 5.3 and 5.4 need to be rewritten as they refer to the previous split document structure.

20 Metadata catalogue: RLS could also be investigated here. Currently there's only gLITE's LFC. But maybe this should be discussed in IGE first.