Cloud resources: Difference between revisions

From Alliance Doc
Jump to navigation Jump to search
No edit summary
No edit summary
Line 8: Line 8:
|-
|-
|  
|  
'''GP1 nodes'''<br/>
'''West GP1 Nodes'''<br/>
{| class="wikitable sortable"
{| class="wikitable sortable"
|-
|-
Line 19: Line 19:
| 2 || 2 x Intel E5-2680 v4 || 128 || 4 x 900GB 10k SAS in RAID5 || 1 x 10GbE || 2 x Tesla K80 || 56
| 2 || 2 x Intel E5-2680 v4 || 128 || 4 x 900GB 10k SAS in RAID5 || 1 x 10GbE || 2 x Tesla K80 || 56
|}
|}
'''West Cloud'''<br/>
'''West Original Nodes'''<br/>
{| class="wikitable sortable"  
{| class="wikitable sortable"  
|-
|-

Revision as of 22:23, 13 October 2016

Other languages:

Parent page: CC-Cloud

Hardware

West Cloud/GP1 East Cloud

West GP1 Nodes

Node Count CPU Memory (GB) Storage Interconnect GPU Total Cores
240 2 x Intel E5-2680 v4 256 4 x 900GB 10k SAS in RAID5 1 x 10GbE N/A 6720
8 2 x Intel E5-2680 v4 512 4 x 900GB 10k SAS in RAID5 2 x 10GbE N/A 224
2 2 x Intel E5-2680 v4 128 4 x 900GB 10k SAS in RAID5 1 x 10GbE 2 x Tesla K80 56

West Original Nodes

Node Count CPU Memory (GB) Storage Interconnect GPU Total Cores
32 2 x E5-2650 v2 CPU's 256 3 x 600GB 10K SAS in RAID0 1 x 10GbE N/A 512
8 2 x E5-2650 v2 CPU's 512 3 x 600GB 10K SAS in RAID1E 1 x 10GbE N/A 128

total compute cores: 7640

200 TB of Ceph storage

Node Count CPU Memory (GB) Storage Interconnect GPU Total Cores
36 2 x E5-2650 v2 128 2x 1TB SATA 7.2K in RAID0 + SSD bcache 1 x 10GbE N/A 576

total compute cores: 576

100 TB of Ceph storage

Software

Compute Canada Cloud OpenStack Platform versions as of 19:36, 6 July 2016 (UTC)

  • East Cloud: Mitaka
  • West Cloud: Kilo

Planned Compute Canada Cloud OpenStack upgrades:

  • East Cloud upgrading to Mitaka July 4th

See the OpenStack releases for a list of all OpenStack versions.

Default Allocation

A cloud account comes with the following default maximums:

  • 2 instances
  • 4 virtual CPUs
  • 15360 MB RAM
  • 1 external IP address
  • 40 GB persistent storage
  • 2 volumes
  • 2 snapshots

These limits are intended to support development, experimentation, and small, temporary batch workloads. Requests for more resources should be submitted through the national Resource Allocation Competition.