Core Services Cumulative Monthly Report, March 2005

• New and improved services
• LIT on-call service.
• Institutional Repository.
• Dec: Marsh, in coordination with Blanco, installed Apache on sambuca to enable custom CGI development for the IR project.
• Jan: Snavely installed the Java CoSign software on sambuca to prepare for integrating DSpace with CoSign for authentication to the institutional repository.
• Mar: Snavely, with assistance from Blanco and Farber, coordinated the design of a development and production environment for DSpace that allows for multiple work areas for developers and/or mantainers and complete version control, including tracking and importing new releases of DSpace from MIT and preservation of local configuration.
• Mar: Snavely configured Java CoSign on sambuca, establishing a proof-of-concept that logic within the DSpace application will be able to trigger CoSign authentication.
  
• Shibboleth.
• Mar: Snavely, working with Doster at ITCS, began the process of establishing a pilot service for access to ProQuest resources.
  
• Infrastructure
• Improvements to storage infrastructure, web server redundancy.
• Nov: Marsh and Snavely selected a high-performance switch to underlie the first storage network, and Snavely placed the order.
• Feb: Marsh installed and configured the new storage server, whisky, and the storage network switch in the Central Campus server environment.
• Mar: Snavely synchronized content from dlps10 to whisky, reactivated mechanisms in the DLXS workflow to automatically synchronize both servers when content or middleware is updated and when page images are loaded, and automated a weekly synchronization as a backup to these routines.
  
• CoSign.
  
• Jan: Snavely installed and configured CoSign on bernie, the new Web Services development server. This will provide an environment for Web Services to migrate web applications that use Library Authentication to begin using CoSign.
• Positioning DLPS as a trusted digital repository (TDR).
• Dec: Snavely, working with Stephenson and Willett, reviewed options for persistent identifier schemes for the library and settled on a preference for CNRI handles. This preference will be reviewed with appropriate colleagues within the library.
• Jan: Snavely installed and configured the Handle Server software on rum and registered the library with CNRI's global registry.
• Mar: Snavely installed and configured a dedicated handle server for DSpace on sambuca with a subordinate naming authority.
• Mar: Snavely, meeting with Stephenson and Willett, discussed approaches for introducing automation into workflows in the Digital Reformatting unit.
  
• Redundant MySQL.
• Mar: March and Snavely discussed configuration plans for redundant MySQL servers, and Marsh placed the order.
  
• DHCP/DNS services.
• Nov: Marsh, in coordination with DSS, rolled out temporary DHCP service on sake after the library's primary DHCP server, stella, failed. Plans are still being developed to move to ITCS-managed DHCP.
  
• Develop an authentication and authorization management system.
• Revisit logging practices.
  
• Production server replacements/upgrades.
• Nov: Marsh installed and configured scotch, a new file server for the Michigan Digitization Project.
• Dec: Marsh and Snavely decided on a storage upgrade configuration for gracie to allow for more free space for Oracle archive logs and disk-to-disk backups, and Marsh placed the order.
• Jan: When load scripts began to warn of decreasing free space, Marsh ordered a storage upgrade for dlps10.
• Jan: Storage purchased for ARC arrived and Marsh and Snavely devised a new storage upgrade plan, diverting the new storage to gracie to accommodate a new, urgent need to do parallel indexing. Marsh began the process of installing and configuring the storage and ordered a new SCSI adapter for ethel.
• Feb: Marsh installed the storage upgrade for Oracle archive logs and disk-to-disk backups on gracie and Snavely migrated the storage into place on the system, thus freeing storage that will be used on ethel. Marsh continued work on the parallel indexing storage upgrade for gracie, eventually involving Sun technical support due to a suspicion of hardware problems.
• Mar: Marsh, working with Sun technical support, resolved all outstanding issues with the new storage, connected the new storage to gracie and the old repurposed storage to ethel, and conducted load testing on both.
  
• Mar: Snavely rebuilt disk configurations on the new storage for ethel and gracie and put both into service.
  
• Formalize system administration infrastructure.
• Replace DLXS statistics.
• Centralize LIT backup.
• Feb: Marsh arranged to acquire two large tape libraries from colleagues at ITCS for potential use in a centralized LIT backup service.
  
• Development server replacements/upgrades.
• Jan: Marsh, working with Steinberg, completed the transition of Web Services to a new Linux-based development server, bernie.
• Jan: Snavely installed Oracle and Library Authentication on bernie.
  
• Jan: Snavely dumped the journal management system database so Sheppard could migrate the service, one of the last fizzie dependencies, to production servers.
  
• Security assessment.
• Workflow, processes, and organizational work
• LIT work requests and scheduling.
• Nov: Snavely recommended the use of FootPrints to LIT managers for handling work requests and scheduling, and suggested a brief refresher session for FootPrints use for the group, to focus on reporting.
• Dec: Snavely made basic configuration tweaks to the LIT Core Services project in FootPrints, and Marsh and Snavely began entering current work in process into the tracking database.
  
• Clearly articulate LIT services and resources.
• Dec: Snavely prepared an overview of Core Services, to be integrated with overviews from other LIT units.
  
• Documentation and standardization of processes.
• Dec: Marsh and Snavely moved several pockets of old documentation into the new structure on the LIT web site, and Snavely did some basic reorganization, simplification, and editing.
  
• Improve coordination of data loading.
• Cost model.
• Technology trends awareness.
• Miscellaneous/Unplanned
• Nov: Marsh continued to migrate more users from the local mail server, rum, to ITCS mail.
• Dec: Marsh migrated all but the final few users to ITCS mail.
• Dec: Snavely configured Samba file sharing on the MDP file server and, in coordination with Gietzen and Potter, finalized the design of scripts that will automate the quality review process.
• Jan: Marsh completed the migration of mail users from the local mail server, run, to ITCS mail.
• Jan: To work around a Solaris system limitation, Snavely compiled and installed a new 64-bit version of Perl on dlps10. After testing, Marsh distributed the new version of Perl to dlps8 and dlps9.
• Jan: Snavely compiled and installed throttling software on dlps8, dlps9, and dlps10 to prevent overuse of the systems by automated downloaders.
  
• Feb: After limited success and several attempts at tuning, Snavely concluded that the throttling software added an unacceptable amount of overhead, removed it, and began seeking other options.
• Feb: Upon learning that the backup service currently used by DLPS servers on north campus will be discontinued, Snavely pursued a preliminary agreement with ITCS for backup services and began preparations for migrating.
• Feb: Marsh investigated a potential problem with memory on gracie, and was advised by Sun that the error was harmless.
• Mar: Snavely completed migrating all DLPS servers to ITCS backup services.
• Mar: Snavely tested download and validation processes for a large batch of material from the Google return server, tweaking existing scripts as needed.
  
• System Performance
  
• Arbor Lakes server environment
• ethel.umdl (ARC server): No down time
• gracie.umdl (Aleph server): Down on Tuesday, March 1 from 11:50am to 1:50pm to resolve a problem with a storage subsystem and on Sunday, March 6 from 8:00am to 12:45pm for work related to storage subsystem upgrades.
  
• hiram.umdl (Zebra server): No down time
  
• tequila.umdl (Metalib and SFX server): No down time
  
• Central Campus server environment
• betty.umdl (MIRLYNWeb server): No down time
  
• coffee.umdl (Oracle server for access control): No down time
  
• gin.umdl (WebCheckout server): No down time
• myrna.umdl (Library Web server): Down briefly on Wednesday, March 30 at approximately 5:15pm due to a disk space shortage.
  
• quik.umdl (Survivors of the Shoah cache server): No down time
• ting.umdl (library proxy server): No down time
  
• North Campus server environment: All services were unavailable from Wednesday, March 16 at 9:30pm to Thursday, March 17 at 5:30am and from Tuesday, March 30 at 9:00pm to Wednesday, March 31 at 8:45am due to network and related problems.
• dlps5.umdl (backup library proxy server): No additional down time
• dlps6.umdl (Oracle server for access control and usage statistics): No additional down time
  
• dlps7.umdl (Numeric Data server): No additional down time
  
• dlps8.umdl (non-public collections): Also down on Tuesday, March 15 from 8:00am to 8:25am for routine maintenance.
  
• dlps9.umdl (Image Services): No additional down time
• dlps10.umdl (primary web server for DLPS and public collections): No additional down time
  

• Data Loading and Archival Statistics