Review of critical area requests

Wednesday 26 Oct 2022, 11:00 → 12:00 Europe/Zurich

513/R-070 - Openlab Space (CERN)

critical area resources.xlsx

Present: Bernd, Tim, Wayne

 

• Highest priority is to reduce the time to recover. This implies
  • systems that allow us to access and recover applications
  • the monitoring of basic infrastructure (power, network)
• Thus, lxtunnel request is approved and is small
• Given 80 cores in total spare,
  • Approve 50 cores for the tomcat services
    • Recommend priority for Adams, Foundation and Phonebook as being the ones we need for the recovery
  • Need to clarify the load balancer quota since LBaaS will be down if physics power is down
• The CephFS move would be recommended if feasible
  • Use of the Hyperconverged cluster to hosts Ceph mons should be evaluated

Actions

 

• Tim to inform the cloud team of the proposal
• Tim to inform Jose, Enrico, Arne re: CephFS to validate the hyperconverged proposal
• Wayne to review the resources for opendcim and TIM to check we've got sufficient
  • If there is a need, these should be considered as highest priority for any remaining capacity
• Tim and Bernd to review critical area resources list with the aim of
  • Finding remaining cores for tomcat
  • Moving towards a cleaner situation for the barn going forward. DFS capacity, for example, would seem to be a worthwhile evaluation.

 

 

 

11:00 → 11:20 Input from resource requests

Current resources in the critical area in attached spreadsheet.

 

1) Y/N for CephFS move to the barn

2) There are currently several outstanding requests for capacity. We need to Y/N them. Cloud team estimate space for 20 4xcore VMs.

 

- 2 cores for lxtunnel (RQF2150048)

- 100 cores for Tomcat (RQF2150308), including LBs

- 48 cores for CephFS mons in the barn (RQF2151464)

 

The details of the Tomcat request is as follows

 

egroup - k8s cluster (1/2pod) + webservice?
Kitry - 1 app VM (application), VM (httpd) appserver/oracleforms11g/kitry_prod
Adams (Apex) - K8s cluster (4 pods), VM (httpd)
Foundation -  K8s cluster (1 pod), VM (httpd)
phonebook - k8s cluster (1 pod)
Qualiac - 2 app VMs, VM (httpd) appserver/qualiac/qualiac_prod
EDH - k8s cluster (2 pods)
Baan/InfoLN - 2 app VMs, VM (httpd) appserver/baan/prod appserver/inforln/inforln_prod
SIR/LMS - 1 app VM, VM (httpd) appserver/lms/lms_prod

VM flavours:
Application VM - m2.2xlarge (8 cores)
apache VM - m2.large, 1 VM can be shared for all applications (4 cores)

to be discussed:
EDMS?
InforEAM
NFS from old IT-DB

day2 - efiles, payroll, OracleHR, Impact are deployed now as follows:
 

efiles, Impact - kubernetes

payroll - VMs

OracleHR - physical machines in physics area