Difference between revisions of "OBCP Informatics Portal"

From xBio:D Wiki
Jump to navigation Jump to search
(Current Configuration)
 
(9 intermediate revisions by the same user not shown)
Line 5: Line 5:
  
 
== General Information and Policies ==
 
== General Information and Policies ==
The OBCP cluster was purchased as part of a loose coalition between the Ohio Biodiversity Conservation Partnership (OBCP), the [http://ael.osu.edu/ Aquatic Ecology Laboratory (AEL)], and the Center for Biodiversity Research & Analysis (CEBRA) and provides computational support for parties associated with these groups. The cluster is designed to facilitate population genetics, ecological modeling, phylogenetic analyses, or any other computationally intensive scientific need for the researchers involved.
+
The OBCP cluster was purchased as part of a loose coalition between the [http://obcp.osu.edu/ Ohio Biodiversity Conservation Partnership (OBCP)], the [http://ael.osu.edu/ Aquatic Ecology Laboratory (AEL)], and the Center for Biodiversity Research & Analysis (CEBRA) and provides computational support for parties associated with these groups. The cluster is designed to facilitate population genetics, ecological modeling, phylogenetic analyses, or any other computationally intensive scientific need for the researchers involved.
  
 
Rather than begin with a restrictive policy towards resource utilization per user, we intend to take a more hands-off approach that assumes that all users will exhibit self restraint towards resource monopolization. Although potentially naive, we do not want to create an unnecessary impediment towards cluster usage that may occur from firm CPU / memory limits, batch job queuing, etc. If you are having difficulties with successfully running a task which you believe is caused by a lack of available cluster resources, please email [mailto:cora.1@osu.edu Joe Cora] and explain your issue. A restrictive policy towards resource usage will be enacted if a proper equilibrium amongst users can not be achieved organically.
 
Rather than begin with a restrictive policy towards resource utilization per user, we intend to take a more hands-off approach that assumes that all users will exhibit self restraint towards resource monopolization. Although potentially naive, we do not want to create an unnecessary impediment towards cluster usage that may occur from firm CPU / memory limits, batch job queuing, etc. If you are having difficulties with successfully running a task which you believe is caused by a lack of available cluster resources, please email [mailto:cora.1@osu.edu Joe Cora] and explain your issue. A restrictive policy towards resource usage will be enacted if a proper equilibrium amongst users can not be achieved organically.
  
 
For a quick insert on the benefits and specifications of OBCP cluster for grants, please refer to the [[OBCP Cluster Description for Grants]] page.
 
For a quick insert on the benefits and specifications of OBCP cluster for grants, please refer to the [[OBCP Cluster Description for Grants]] page.
 +
 +
 +
When research produced with the help of the cluster is presented in a talk, poster or publication, please use the following text or an approximation: "''This work was supported by computation resources of the Ohio Biodiversity Conservation Partnership (OBCP) cluster''".
  
  
Line 15: Line 18:
  
 
=== Storage Policy ===
 
=== Storage Policy ===
'''PLEASE READ''': Each user has a directory on the OBCP network attached storage (NAS) device that allows unlimited, within reason, storage space. Please store all of your data files in this directory. The data directory, named ''data'', can be found in the home folder of each user. Type ''cd $HOME/data'' to access the folder from whichever your current working directory may be. You are limited to 1GB of storage on the OBCP OS disk, and the operating system will remove or disallow files that exceed this limit.
+
'''''PLEASE READ''''': Each user has a directory on the OBCP network attached storage (NAS) device that allows unlimited, within reason, storage space. Please store all of your data files in this directory. The data directory, named ''data'', can be found in the home folder of each user. Type ''cd $HOME/data'' to access the folder from whichever your current working directory may be. You are limited to 1GB of storage on the OBCP OS disk, and the operating system will remove or disallow files that exceed this limit.
  
  
Line 30: Line 33:
 
* '''Server'''
 
* '''Server'''
 
** ''Model'': Dell PowerEdge R820
 
** ''Model'': Dell PowerEdge R820
** ''OS'': Ubuntu 12.04 64-bit
+
** ''OS'': Ubuntu 14.04 LTS 64-bit
 
** ''CPU'': 4 x Intel Xeon E5-4650L (8 cores, 16 threads) -> 64 available pseudo-processors
 
** ''CPU'': 4 x Intel Xeon E5-4650L (8 cores, 16 threads) -> 64 available pseudo-processors
 
** ''RAM'': 256GB (16 x 16GB) (upgradeable to 768GB)
 
** ''RAM'': 256GB (16 x 16GB) (upgradeable to 768GB)
Line 39: Line 42:
 
** ''CPU'': Intel Xeon E5620 (4 cores)
 
** ''CPU'': Intel Xeon E5620 (4 cores)
 
** ''RAM'': 12GB
 
** ''RAM'': 12GB
** ''HD'': 6TB (RAID6 - 4 x 3TB) (upgradeable to 30GB)
+
** ''HD'': 25.1TB (RAID6 - 11 x 3TB) (upgradeable to 28TB)
  
 
=== Upgrade Potential ===
 
=== Upgrade Potential ===
Line 53: Line 56:
  
 
==== Storage Capacity Upgrade ====
 
==== Storage Capacity Upgrade ====
Adding additional storage capacity to the NAS is easy. The NAS is setup to accept the failure of two hard disks at the same time without losing any data from storage. This requires a large amount of storage overhead for limited disk arrays but allows subsequent disks added to contribute almost all of its space towards the effective capacity. This means that by adding a single, 3TB disk to the current setup the total effective storage space will also increase by 3TB, and so on for additional disks. Below are the specifications for a hard disk:
+
Adding additional storage capacity to the NAS is easy. The NAS is setup to accept the failure of two hard disks at the same time without losing any data from storage. This requires a large amount of storage overhead for limited disk arrays but allows subsequent disks added to contribute almost all of its space towards the effective capacity. This means that by adding a single, 3TB disk to the current setup the total effective storage space will also increase by 3TB, and so on for additional disks. Below are the specifications for a hard disk and a required hard disk caddy:
  
 
* '''Hard Disk'''
 
* '''Hard Disk'''
Line 59: Line 62:
 
** ''Model #'': ST33000650SS
 
** ''Model #'': ST33000650SS
 
** ''Price'': ~$360 (as of 06-Sep-2012)
 
** ''Price'': ~$360 (as of 06-Sep-2012)
 +
* '''Hard Disk Tray/Caddy'''
 +
** ''Model'': Dell NX3100 Hard Disk Tray/Caddy
 +
** ''Model #'': F238F
 +
** ''Price'': ~$11-50 (as of 22-Nov-2013)
  
  

Latest revision as of 15:08, 15 April 2015

Ohio Biodiversity Conservation Partnership (OBCP) Informatics Wiki

This section of the wiki is designed to instruct and share with users the features of the Ohio Biodiversity Conservation Partnership (OBCP) Informatics Infrastructure housed within the C.A. Triplehorn Insect Collection Computing Rack. The rack is administered by Joe Cora, so direct any software or hardware queries to him.


General Information and Policies

The OBCP cluster was purchased as part of a loose coalition between the Ohio Biodiversity Conservation Partnership (OBCP), the Aquatic Ecology Laboratory (AEL), and the Center for Biodiversity Research & Analysis (CEBRA) and provides computational support for parties associated with these groups. The cluster is designed to facilitate population genetics, ecological modeling, phylogenetic analyses, or any other computationally intensive scientific need for the researchers involved.

Rather than begin with a restrictive policy towards resource utilization per user, we intend to take a more hands-off approach that assumes that all users will exhibit self restraint towards resource monopolization. Although potentially naive, we do not want to create an unnecessary impediment towards cluster usage that may occur from firm CPU / memory limits, batch job queuing, etc. If you are having difficulties with successfully running a task which you believe is caused by a lack of available cluster resources, please email Joe Cora and explain your issue. A restrictive policy towards resource usage will be enacted if a proper equilibrium amongst users can not be achieved organically.

For a quick insert on the benefits and specifications of OBCP cluster for grants, please refer to the OBCP Cluster Description for Grants page.


When research produced with the help of the cluster is presented in a talk, poster or publication, please use the following text or an approximation: "This work was supported by computation resources of the Ohio Biodiversity Conservation Partnership (OBCP) cluster".


The domain name for the cluster is obcp-server.biosci.ohio-state.edu.

Storage Policy

PLEASE READ: Each user has a directory on the OBCP network attached storage (NAS) device that allows unlimited, within reason, storage space. Please store all of your data files in this directory. The data directory, named data, can be found in the home folder of each user. Type cd $HOME/data to access the folder from whichever your current working directory may be. You are limited to 1GB of storage on the OBCP OS disk, and the operating system will remove or disallow files that exceed this limit.


Accessing the Cluster

For some information on accessing the OBCP cluster, visit the Accessing the OBCP Cluster page.


Multiple Processor Runs on the Cluster

For some information on running parallel processes on the OBCP cluster, visit the Multiple Processor Runs on the OBCP Cluster page.


Hardware

Current Configuration

  • Server
    • Model: Dell PowerEdge R820
    • OS: Ubuntu 14.04 LTS 64-bit
    • CPU: 4 x Intel Xeon E5-4650L (8 cores, 16 threads) -> 64 available pseudo-processors
    • RAM: 256GB (16 x 16GB) (upgradeable to 768GB)
    • HD: 146GB (RAID0)
  • Network Attached Storage (NAS) Unit
    • Model: Dell PowerVault NX3100
    • OS: Windows Storage Server 2008 R2 x64
    • CPU: Intel Xeon E5620 (4 cores)
    • RAM: 12GB
    • HD: 25.1TB (RAID6 - 11 x 3TB) (upgradeable to 28TB)

Upgrade Potential

The initial configuration of the OBCP cluster was designed to allow hardware upgrades to both RAM and hard disk storage capacity. In order for an upgrade to occur, specific equipment must be used that matches the current configuration.

RAM Upgrade

RAM upgrades must be made in batches of four to accommodate a new RAM chip for each processor, thus limiting the number of upgrades to two separate events. Below are the specifications for a RAM chip:

  • RAM
    • Model: Samsung 16GB (PC3-12800) DDR3-1600MHz, ECC Registered DIMM, Dual Rank
    • Model #: M393B2G70BH0-CK0
    • Price: ~$386 (as of 06-Sep-2012)

Storage Capacity Upgrade

Adding additional storage capacity to the NAS is easy. The NAS is setup to accept the failure of two hard disks at the same time without losing any data from storage. This requires a large amount of storage overhead for limited disk arrays but allows subsequent disks added to contribute almost all of its space towards the effective capacity. This means that by adding a single, 3TB disk to the current setup the total effective storage space will also increase by 3TB, and so on for additional disks. Below are the specifications for a hard disk and a required hard disk caddy:

  • Hard Disk
    • Model: Seagate 3TB 7200RPM 64MB ES.2 SAS
    • Model #: ST33000650SS
    • Price: ~$360 (as of 06-Sep-2012)
  • Hard Disk Tray/Caddy
    • Model: Dell NX3100 Hard Disk Tray/Caddy
    • Model #: F238F
    • Price: ~$11-50 (as of 22-Nov-2013)


Software

For a list of the available software present on the OBCP cluster, visit the Available Software on the OBCP Cluster page.