Strumenti Utente

Strumenti Sito


oph:cluster:monitoring

Differenze

Queste sono le differenze tra la revisione selezionata e la versione attuale della pagina.

Link a questa pagina di confronto

Entrambe le parti precedenti la revisioneRevisione precedente
Prossima revisione
Revisione precedente
oph:cluster:monitoring [2023/09/28 12:52] marco.baldi5@unibo.itoph:cluster:monitoring [2023/09/29 08:36] (versione attuale) marco.baldi5@unibo.it
Linea 1: Linea 1:
 ===== Cluster Monitoring ===== ===== Cluster Monitoring =====
  
-This page show the real-time monitoring of the usage of the OPH cluster resources for the current day (from 00h00:00 to 23h59:00). The two plots show the usage of the whole cluster (top) and of the Matrix nodes (bottom) as the fraction of available CPUS in different categories: +This page shows the real-time monitoring of the usage of the OPH cluster resources for the current day (from 00h00:00 to 23h59:00). The two plots show the usage of **the whole cluster (top)** and of **the Matrix nodes (bottom)** as the fraction of allocated/available CPUs in different categories:
-  * DIFA: all the available CPUs  +
-  * DIFA: resources available to all cluster users through general access based on standard accounts +
-  * Resv: resources reserved to specific projects and accessible to a limited set of users/groups through additional special accounts +
-  * Overcommit: fraction of CPUs that are requested by queueing jobs but cannot be allocated due to limitations on the type or size of requested resources +
-  * Blocked: fraction of CPUs that are requested by queueing jobs and that could be allocated on idle reserved nodes that are not accessible by the jobs+
  
 +  * **Total**: all the available CPUs 
 +  * **DIFA**: resources available to all cluster users through general access based on standard accounts
 +  * **Resv**: resources reserved to specific projects and accessible to a limited set of users/groups through additional special accounts
 +  * **Overcommit**: fraction of CPUs that are requested by queueing jobs but cannot be allocated due to limitations on the type or size of requested resources
 +  * **Blocked**: fraction of CPUs that are requested by queueing jobs and that could be allocated on idle reserved nodes that are not accessible by the jobs
 +
 +{{https://apps.difa.unibo.it/files/people/Str957-cluster/oph/Cluster/monitoring_now.jpg?nocache}}
 +
 +{{https://apps.difa.unibo.it/files/people/Str957-cluster/oph/Cluster/monitoring_matrix_now.jpg?nocache}}
  
oph/cluster/monitoring.1695905522.txt.gz · Ultima modifica: 2023/09/28 12:52 da marco.baldi5@unibo.it

Donate Powered by PHP Valid HTML5 Valid CSS Driven by DokuWiki