ORC === - 130TB with RAID6 installed - mounted drives only showing 500GB VMWare limitations: file size limitation of 2TB - one physical adapter connected to iSCSI (2 ports through switch to drive arrays) - (perhaps) Need to reconfigure to see the full partition - Need responsive admin support at ORC. Action: DAVE - check with Bruce to try and get admin support a little more responsive Aaron Lesch Leslie Ault Larry Jennings According to the documentation, here is the PCI slot configuration: 1 - PCIe slot 1 PCI Express (Generation 2) x8 link expansion slot (24.13 cm [9.5"] length). 2 - PCIe slot 2 PCI Express (Generation 2) x4 link expansion slot (low-profile 24.13 cm [9.5"] maximum length, with a standard height bracket). 3 - PCIe slot 3 PCI Express (Generation 2) x8 link expansion slot (low-profile 24.13 cm [9.5"] length). 4 - PCIe slot 4 PCI Express (Generation 2) x8 link expansion slot (low-profile 24.13 cm [9.5"] length). 5 - PCIe slot 5 PCI Express (Generation 2) x8 link expansion slot (24.13 cm [9.5"] length). 6 - PCIe slot 6 PCI Express (Generation 2) x8 link expansion slot (24.13 cm [9.5"] length). The current dual-port iSCSI controller cards occupy slots 5 and 6. Slots 1-4 are available, however they are all low profile except for slot 1. Also, slot 2 is x4 where all the others are x8. I don't know what form factor the 10G network cards are, but it looks like slots 1, 5, and 6 are the only ones with x8 speed that will accept a full height card. It looks like slots 3 and 4 will be your only option to keep the existing iSCSI cards and have 2 "like" 10G network cards (assuming they are available in low profile versions). If you want 4 connections to the iSCSI controllers as exist now, then each card will need 2 10G ports. - 3 hosts - each host 2 processors, each with 12 cores, 2.5GHz (24 cores / host) - 131GB RAM per host - 140GB drive per host (130 formatted) - 4 x 1GB NICs - 1 connected to VMware kernel management network - 1 vmware traffic - 1 iSCSI-1 - 1 iSCSI-2 - 10GB NICs on order (4 ports per adapter)?? - 2 x storage arrays. Each array consists of: - 1 x MD3600i - 7 x MD1200i - 96 2TB drives per array - 16disk groups - 4 hot spares per "rack" - VMWare high availability license - 3 x powerconnect 8024 - UCSB ==== - Hardware powered up, continuing installation - hardware configuration pretty much done - install host OSs - setup storage 4 10GB NIC ports per server 4 1GB ports per server - 4 switches configured to operate as two (one redundant) HOST 1 10 GB nic 1 - switch 1 - iSCSI 1 10 GB nic 2 - switch 2 - iSCSI 1 10 GB nic 3 - switch 1 - WAN (1GB) 10 GB nic 4 - switch 2 - WAN (1GB) ilo to management switch (1GB) four separate drive arrays 120 TB raw storage / array. Four arrays Hardware List ==At NCEAS== 1x Dell Server (PowerEdge R815): - 16 AMD Opteron Cores - 64 GB memory - 2.4 TB local storage (18 free 2.5" drive bays for expansion) -- 4x 600 GB 15K RAID10 internal drives -- 4x 600 GB 10K RAID10 external SAS array (Dell PowerVault MD3220, 20 empty bays) ==On Campus== 4x HP Servers (DL385 G7), each with: - 24 AMD Opteron Cores (6180SE, 2.5GHz) - 128 GB memory - 146 GB local storage (2x 146GB 10K SAS in RAID1) - 4x 10-GbE SFP+ ports (2x Dual-port PCIe NICs) - 4x gigabit ports (internal) - iLO 3 Advanced management port 4x 10-GbE iSCSI Storage Arrays (P2000), each with - 60 2TB drives per array - 1 P2000 controller, 4 P2000 enclosures (12 drives each) - expected 100 TB expected usable space - 120 TB raw - mountable on any/all servers 4x 10-GbE Switches (HP Procurve A5820) - configured as two virtual switches with failover - two dedicated to networking - two dedicated to iSCSI 1x Gigabit Switch (HP - for the server/array management network Need to know: - number of hosts - cores per host - NICs per host