[Previous] [Contents] [Next]

Disaster Planning and Your INNOPAC: Theory and Reality

Lynne Lysiak, Appalachian State University
Carmel Bush, Colorado State University
Nora Copeland, Colorado State University
Tom Jacobson, Innovative Interfaces, Inc.

Carmel Bush, a member of the library’s Disaster Recover Team (DRT), presented the situation and described what improvements were made to the disaster plan. On July 28, 1997, a wall of water from torrential downpours resulted in 8½ feet of water flooding the lower level of the library at Colorado State University.

Four hundred thirty-three thousand volumes of both bound journals and monographs were affected with estimated damage was $40-$50 million. Cable tracks were in the ceiling, and the flood came within inches of reaching these. All underwater connections had to be reworked, including hubs and switches. The INNOPAC was housed on the first level and escaped damage.

Personal safety is always the first consideration in a disaster so the building was evacuated. The second consideration is powering-off the computers. The DRT member for technology, a campus security officer and a facilities person did this. Limited backup power supplies were available, and they were in use because the electricity had already been cut off.

A viable disaster plan depends on a library’s own risk assessment. This disaster was beyond the scope of the original plan. That document was written for a total destruction; there was no pre-arranged space to move into during the recovery process, nor was there a plan for an alternate site for use during the transitional, recovery process. The plan did not allow for a flood that produced floating materials, making it impossible to remove volumes from shelves for packing. Additionally, user needs continued and immediate consideration was given to providing INNOPAC and interlibrary loan service to users. Two computers were wired from the engineering main site, and service was restored after two days. Computer use was key to recovery; coaxial network wiring had to be run to the recovery depository very quickly.

Communication covered everything from staff members to consortia to vendors. Cell phones were used on campus, and late-night calls were made from homes. Getting the word out to the user community became important. The Colorado Alliance for Research Libraries provided great assistance, and OCLC turned the CSU library into a non-lending library. Vendors were notified and services were suspended. Tables of Contents services for example, were not needed, since there was no collection.

The environment had to be stabilized quickly. The water had been contaminated with sewage and other debris. Pumping the water took 1½ days, followed by thorough cleaning. The HVAC system was destroyed, and replacement was one of the first contracts allowed. This contractor remained on the site throughout the recovery process. Environmental perceptions were very important to many staff; safety and testing data was available for everyone who asked.

Another inadequacy of the disaster plan was not allowing for the effect of other buildings on campus being destroyed. One of the 30 buildings involved at CSU was the police department. Securing the library building became a nightmare. Staff was positioned at only the front door; the gaping hole at the back of the building was inadvertently not guarded. The number of workers constantly in the area made guarding property difficult. A recommendation would be to hire a separate security firm for this policing service.

Missing from the disaster plan were the applications necessary to recover from a disaster. Most library disaster plans intimate that control of the recovery remains in the library, rather than being dictated by or influenced by a campus disaster plan or the insurance provider. It was necessary to demonstrate the needs of the library’s recovery. The Library of Congress served as a FEMA consultant, and provided great assistance in this area. A processing plant was developed for the function of database management, inspections, paper repair, inventory, quality control and shelving. The overall design of the plant took several months while the university negotiated the contract with the emergency management facility company. It was anticipated that 7,000 volumes per week would be processed coming back from the freeze-dry facility. A separate database was developed for the materials; Innovative assisted in how the data would be prepared and output, and how the INNOPAC would be updated. Gifts began to arrive during this period, and it was necessary to change the plan. INNOPAC was used to verify what was owned and to code the records. Additional staff was necessary for this processing. Records of staff time are important, as they may be reimbursable through insurance.

The insurance company wanted to use a salvage company for recovery of hardware, but materials returned from this company were of poor quality. Much time was spent defending the library’s position that computers should be replaced, not refurbished. A cross-campus agreement of refurbishment or replacement would have been helpful. Maintenance contracts may release the provider from continuing service if the equipment has been refurbished.

Nora Copeland discussed the role and use of the INNOPAC in inventory control, statistics, and information retrieval through the Create List function. Her defining title “You get what you ask for (always); you get what you want (maybe)” refers to INNOPAC always retrieving the information you search for based on fields and values. This may not be the information you want.

Know the history of the library’s database: have all volumes been barcoded; do all have item records in the database; are location codes correct, keeping in mind holdings which have been shifted to other locations. Circulation statistics that have been zeroed out will not provide a true picture for determining which titles are high use. You should also know what kind of records are in the database and what cataloging practices have been used.

Know what codes and fields have been used in records: what data are in the fixed fields, item records, bibliographic records, and checkin records; where is the call number stored. Understand the structure of INNOPAC: how records exist with the bibliographic record; what strategies will retrieve the most appropriate type of list to fulfill the objective.

Innovative provided immediate assistance by loaning the library larger file sizes for Create Lists. Innovative also discussed how to use the system to the library’s advantage, and how to create the necessary codes. The 9xx MARC fields were used to identify flooded items and disposition. This helped not only in identifying the flood items, but was useful for insurance purposes. When the recovery was complete and the codes were no longer needed, Innovative deleted the field tags instead of having the library do it.

One application was to combine the Mat type with the YTD circulation >3 and Status as available (not checked out) at the time of the flood. This identified the heavily used titles for immediate replacement. When replaced, the original item was designated as a withdrawn item replaced by an exact purchase item. Statistics can be used to determine how many items have been purchased and replaced. Flexibility should be built in to record codes; be creative in combining record types and fields.

Temporary workers not versed in database management or library experience staffed the processing center. It was reluctantly decided that temporary staff could update the item records. Detailed manuals were created, and training included MARC fields and searching for bibliographic and item records.

Staff was taught quality control and how to add 975 and 990 fields. The criteria and guidelines were extremely important. Mistakes were less than 2%.

Based on the 9xx fields remaining in the records, longitudinal studies will be conducted on condition of items returned from freeze-drying and on the shelves, to monitor for mold or odor.

Innovative provided exactly the type of response one would expect from a vendor. It has been a working partnership.

Tom Jacobson described Innovative’s approach to this disaster as different from other disasters: the INNOPAC was intact. This changed Innovative’s approach to more of a “How can we help you” from a “system down scramble” approach. In either case, Innovative’s recovery emphasis involves looking at the system in two ways: the database, unique to the library; and the hardware, readily replaced, generally with the next generation. Notify Innovative as quickly as possible about the disaster, even if this is only a call to the Help Desk. The following highlight his handout “Disaster Planning for Automated Systems.”

The vendor’s recommended procedures for the protection of the data should be audited before a disaster occurs. See what really happens versus what the procedures say. Ascertain who is responsible for hardware; how is leased hardware versus purchased hardware handled; and the differences between actions for peripherals and the CPU.

Acts of God (force majuere) are not covered by Innovative’s maintenance agreement. When writing the plan, the library must have insurance to cover the INNOPAC. Innovative will remove refurbished peripherals from the maintenance agreement. The CPU replacement will need to be negotiated.

Do not underestimate cost estimates for insurance purposes; include time and material charges for reinstalling equipment and restoring data. Count every penny, including internal costs such as temporary workers.

Restoring the database is similar to the original load: the library gets the data to a certain level, and then Innovative takes over.

Recorded by: Sheryl Williams, University of Nebraska Medical Center

[Previous] [Contents] [Next]