Review of the OSG Critical Probes used for Availability Metrics Calculations

OSG Services

The only authoritative source of information for SAM in terms of OSG services for OPS vo is the OSG interoperability monitoring list, which is accessible through this URL: http://oim.grid.iu.edu/publisher/get_osg_interop_monitoring_list.php

This means that SAM will ONLY monitor those services under OPS credentials.

At present time, the service names that SAM expects to find in that URL are the ones listed in the first column of this table:

Service name in OSG URL	Service name in SAM	Service abbreviation in SAM
OSG-CE	OSG Computing Element	OSGCE
OSG-SRMv1	OSG SRM version 1	OSGSRMv1
OSG-SRMv2	OSG SRM version 2	OSGSRMv2
OSG-BestmanXrootd	OSG BestmanXrootd	OSGBestm
OSG-GridFtp	OSG GridFtp	OSGGdFTP

List of OSG probes

The OSG probes available in the RSV (Resource & Service Validation) project are listed here:

http://rsv.grid.iu.edu/documentation/help/

Equivalence between EGEE and OSG critical tests

Computing Element (CE):

• CE-sft-caver is equivalent to org.osg.certificates.cacert-expiry
• CE-sft-csh is equivalent to ==
• CE-sft-lcg-rm is equivalent to ==
• CE-sft-job is included in ==
• CE-sft-softver is equivalent to org.osg.general.osg-version.txt
• CE-host-cert-valid This one does not apply for OSG (no way to check)
• CE-sft-brokerinfo This one does not apply for OSG (no broker)
• This is not available in EGEE; in OSG org.osg.general.ping-host.txt
• This is not available in EGEE; in OSG org.osg.globus.gram-authentication

Storage Element (SE):

• These do not apply to OSG as they only have SRMs
• SE-lcg-cr
• SE-lcg-cp
• SE-lcg-del

SRM (SRM):

• SRM-put is equivalent to org.osg.srm.srmcp-readwrite
• SRM-get may be equivalent to org.osg.srm.srmcp-readwrite
• SRM-advisory-delete may be equivalent to org.osg.srm.srmcp-readwrite
• This one does not apply to EGEE; in OSG org.osg.srm.srmping

Site BDII (sBDII):

• Does not apply as OSG only uses a global BDII. A future release of RSV will include Global BDII tests.
• sBDII-sanity
• sBDII-performance

Current Critical Tests for OSG defined in SAM (These will be critical through at least June 30, 2008)

• org.osg.certificates.crl-expiry
  • org.osg.certificates.crl-expiry: Checks for expired and expiring-soon CAcerts/CRLs on remote host. Stages worker script defined in OSG_RSV_Probe_Base (or as provided using commandline argument) to remote host; and then checks the *.0 or .r0 files in $OSG_LOCATION/globus/share/certificates/ depending on -m value.
• org.osg.general.osg-directories-CE-permissions
  • org.osg.general.osg-directories-CE-permissions: Checks if permissions are set correctly on important user-accesible OSG directories defined by various common OSG environment variables.
• org.osg.general.osg-version
  • org.osg.general.osg-version: Reports the version of OSG CE running on remote host.
• org.osg.globus.gridftp-simple (remove for proposed status, will be critical for resources with service type OSGGdFTP)
  • org.osg.globus.gridftp-simple: Checks if OSG-GridFtp works to and from remote host. A test file is gridftp'ed to remote host, and back. Then the files are checked for consistency.

Proposed Critical Tests for OSG-CE

• org.osg.general.ping-host.txt
• org.osg.globus.gram-authentication
• org.osg.general.osg-version
• org.osg.certificates.cacert-expiry
• org.osg.certificates.crl-expiry
• org.osg.general.osg-directories-CE-permissions
• July 1st: org.osg.batch.jobmanager-default-status.txt

Proposed Critical Tests for OSG-SRMv1/SRMv2/BestmanXrootd

• org.osg.srm.srmping
• org.osg.srm.srmcp-readwrite

Proposed Critical Tests for OSG-GridFtp

• org.osg.globus.gridftp-simple

-- DavidCollados - 03 Dec 2007