Difference between revisions of "SAM Tests"
Line 60: | Line 60: | ||
The procedure for adding a new probe (PROC06) can be found [[PROC06|here]]. | The procedure for adding a new probe (PROC06) can be found [[PROC06|here]]. | ||
The list of tests can be found [[Operations | The list of tests can be found [[Operations SAM tests|here]]. | ||
== Availability tests == | == Availability tests == |
Revision as of 10:51, 19 October 2012
Main | EGI.eu operations services | Support | Documentation | Tools | Activities | Performance | Technology | Catch-all Services | Resource Allocation | Security |
Tools menu: | • Main page | • Instructions for developers | • AAI Proxy | • Accounting Portal | • Accounting Repository | • AppDB | • ARGO | • GGUS | • GOCDB |
• Message brokers | • Licenses | • OTAGs | • Operations Portal | • Perun | • EGI Collaboration tools | • LToS | • EGI Workload Manager |
Terminology
SAM Test
Test is a procedure which checks specific functionality of a given service, i.e. single measurement (e.g. org.bdii.Freshness, hr.srce.RGMA-CertLifetime). Tests are executed on NGI/ROC SAM instances.
A SAM test is:
- an OPERATIONS test which raises alarms in the Operations Portal (see list of OPERATIONS tests)
- an AVAILABILITY test whose result is used for Resource Centre availability calculation by ACE (see the ROC_CRITICAL profle)
NOTE: CRITICAL is used ONLY to refer to one of the possible results returned by a Nagios probe.
Probe
Probe is code which implements single or multiple tests. See list of SAM Released Probes
Metric
Metric is a synonym for tests used in the development documentation. In operations document "test" is the reference term to be used.
POEM Profile
A POEM is a profile as a triple of (vo, atp_groups and metric instances) where:
- set of atp_groups is a set of service instances defined in the Aggregated Topology Provider via a VO feed, e.g. LHCb_Site LCG.CNAF-T2.it= (service_instance1, service_instance2, etc.)
- set of (service_flavor, metric, fqan) tuples
- metric is fully qualified name of the metric (e.g. hr.srce.SRM2-CertLifetime)
- service_flavor is taken from ATP (e.g. CE, SRM, etc.)
- FQAN - voms role to use for the tests (e.g. /Role=lcgadm) - as metric1 is run with fqan1, metric1 with fqan2, etc.
- VO: is the name of the VO
In other words, the profile is a cartesian product of service groups and metrics, plus VO (read more).
Profiles
- POEM profiles (OPS VO)
- All POEM profiles (OPS VO)
Availability and reliability profile
(Courtesy of P. Andrade, CERN) |
Availability and Reliability Profiles are a collection of metrics/services defined for VOs (multiple profiles per VO). Each profile defines its computation algorithm. Metrics can be in different levels such as crtical, non-critcal etc. |
SAM tests
Tests on NGI/ROC SAM instances are the one which frameworks includes in the SAM configuration. In addition SAM admins can add their own probes to these instances.
SAM teams proposes addition of new probes. The addition of probes is part of SAM release and thus part of the staged rollout. It was agreed that prior to release new list of probes will be briefly presented at the OMB meeting. Probes which perform internal components of SAM are not presented at OMB.
The list of tests included in the SAM release can be found here. Lists of tests to be included are here.
List of new MW related tests: MW SAM tests.
Operations tests
Tests on Operations Portal are the ones used for raising alarms for ROD and COD teams. Operations portal does not execute these tests, but receives alarms from NGI/ROC SAM instances. Operations Portal contains list of the probes used for alarms and others are filtered.
The procedure for adding a new probe (PROC06) can be found here.
The list of tests can be found here.
Availability tests
Set of tests used for calculating availability and reliability of sites and services. The A/R calculation is related to the OLA. As in case of Operations Portal, availability calculation component receives results from NGI/ROC SAM instances.
TSA1.8 proposes a change in avail calculation (which probe results count in it) and the OMB approves.
The list of tests can be found here.