Jornada Basin SiteBytes 2010

2010 Jornada Basin LTER Site Byte

The last year has been very busy. The Jornada Basin LTER (JRN) had its mid-term site review last year. We also virtualized our server farm to prepare for making them highly available using free enterprise class software. We have more than doubled our storage capacity in the last year within our storage area network (SAN) to meet the needs of Jornada research efforts and collaborations. During the process, we migrated and upgraded server operating systems, server platforms (32-bit to 64-bit), server applications, and data stores for all production servers. We also used 2 older servers to deploy virtual servers to support site and research project development efforts.

Server Purchases
The Jornada Experimental Range (JER, USDA ARS) purchased 2 new Dell PowerEdge R710 servers in the last year or so. The Jornada Basin LTER (JRN) is currently purchasing another R710 server using 2010 LTER Site Supplement funds. The R710 servers are used to support all Jornada production virtual servers. 2 older PowerEdge 2950 servers are being used to support virtual servers used for development.

We are using Citrix XenServer software for server virtualization. XenServer is a free enterprise-class virtualization hypervisor. Virtualization reduces administration and maintenance costs of our server farm and allows our servers to be highly available and load balanced to improve performance.

In addition, JRN and JER are cost sharing the purchase of Citrix XenServer Essentials Enterprise for 5 servers to allow automated load balancing and high availability of virtual servers running within 2 XenServer resource pools; production (3 - R710 servers) and development (2 - PE2950 servers).

Storage Expansion
JRN purchased a new IceWeb TE444 SAN storage enclosure half-populated with 24 TB of storage last year. In addition, JRN is currently purchasing hard drives to completely populate the older TE444 storage enclosure with site supplement funds to provide an additional 18 TB of storage. The SAN provides storage for our virtual servers and will provide redundant storage once the new hard drives arrive.

Server Virtualization, Migrations, and Upgrades
All Jornada servers were virtualized within a production resource pool using Citrix XenServer this year, with the exception of 1 physical Active Directory server. The following upgrades (including system and application patches) and migrations were performed during the virtualization process.
• Services, imagery, data, network shares, and permissions were migrated from Novell OES2 to Windows Server 2008 R2
• Operating systems were migrated from 32-bit to 64-bit along with appropriate software applications
• Databases were upgraded from Microsoft SQL Server 2005 to 2008 R2
• All Windows servers were upgraded from Windows Server 2003 to 2008 R2 during virtualization
• 2 Active directory servers (physical, virtual) were updated from Windows Server 2003 to 2008 R2
• The mail server was upgraded from Novell GroupWise 7 to 8 and the operating system was migrated from NetWare 6.5 to Windows Server 2003 and mail accounts were migrated

Novell cluster services (NCS) were leveraged to allow the websites and FTP sites to failover and be hosted by the file server while the web server was added to the production resource pool and the web and FTP sites were virtualized.

The GIS server was configured as virtual host server to support a new development resource pool. The development resource pool contains virtual machines running Windows Server 2008 R2. The virtual servers that were created within the development resource pool support active directory, GIS, database, file, image, and web services needed to support software integration and development and prototyping within a separate environment from our production environment. The Jornada can now quickly clone and export virtual servers from the development resource pool to the production resource pool without impacting services running in either pool. As we complete migration of data from the file server, it will be added to the development resource pool. The development resource pool is being licensed for high availability. Older servers will be used to support development efforts and science projects until ready for deployment in the production resource pool. The file server will be added to the development pool to provide failover and automated load balancing capabilities for the development pool.

LTER Unit Registry Implementation
The Unit Registry Working Group funded Ken Ramsey and Justin Jensen (student programmer) to travel to San Diego in August to implement the LTER Unit Registry into the Jornada Information Management System (JIMS). The units table in our SQL Server database is now populated and synchronized from the LTER Unit Registry with units, metadata, and XML snippets used to create EML custom units for Jornada metadata. Our existing documentation forms (PHP pages) were also modified to allow use of the unit registry. The last step will be to update our XSLT transformations in the Spring Semester to use the custom unit STMML snippets in our units table to create level 5 EML for all Jornada datasets.

Discussion questions:

Given limited resources, how do you prioritize data contributions to the NIS?
This has not been determined as the new NIS has not been developed. We are planning towards integrating JIMS into the new NIS as it is developed. To that end, we are focused on populating all research data into the JIMS relational databases and to generate level 5 EML for all datasets stored in the relational databases. We are also implementing current and evolving best practices from the LTER IMC into JIMS. I answered the following questions in relation to JIMS, not the LTER NIS.

Do you give a lot of attention to a few datasets or limited attention to a lot? We do both. Mostly give a lot of attention to a few datasets. More attention is typically given to new datasets and to updating ongoing datasets. However, we have and continue to update all metadata in our relational databases. In the coming year, we plan on focusing on populating data within the relational database and deriving EML level 5 for all Jornada datasets. This will require verifying all metadata and performing quality assurance data.

Do you deliver derived or raw data?
Both.

Do you communicate with your PIs regarding prioritization of data?
Yes, but the level of communications will ramp up in the next year as data is being documented and populated into the databases.