Endeavor Consortium Work Group Minutes

Tari Keller, Dave Block, Kelly Vickery, Todd King, Margaret Foote, Elsie Pritchard, Gary Flanagan, Linda Voyles, Nelda Sims, Mark Paul, Paul Fuller, Michael Stapleton, Enid Wohlstein, Tanzi Merrit, Rose Davis, Charlene Davis, John Detwiler, Roberta Kirby, Cathy Reilender, Perry Bratcher

Endeavor Update

The EISI move to their new corporate offices has been completed. There is a new phone system and you need to listen to the opening message awhile to discover how to navigate to Technical Support. As of last week there were about 60 EISI AIX customers that have upgraded to 2001.2. Unicode Alpha testing for final data conversion process has been completed with L.C., and they expect to announce the start of Solaris beta testing for Unicode next week. Programming for the release following Unicode has started. A number of improvements (patron purge, bulk loading improvements, acquisitions work) are being investigated and we should check the release notes as they become available.

EISI staffing update: Liz Best is currently in charge of the “Library Support Team Leader”. Jenny Forbes has rejoined EISI training support. EISI is creating a new position to coordinate bug reports and other information for software in the Early Release phase. Don Muccino came on board as V.P. of Customer Support Services in June.

Linda did a demonstration of the new incident report tracking web site at http://support.endinfosys.com/cgi-bin/cust/incident_report.cgi .

Upgrade Evaluation

Dave Block concurred with the comments that Keith Stevenson made in his email to the listserv. From the server administrator perspective the upgrade process was not pretty. They ran the upgrade on the test server and everything appeared to operate as expected. When they upgraded the production servers many things went wrong. A big complaint is that the upgrade instructions did not match the experience; there were many times that the instructions advised to run a command, but when the command was entered it would not operate as expected—things were not running as advertised. Dave speculated that the instructions might have been written for Solaris, and therefore, many commands did not pertain to AIX—many things in the upgrade documentation were wrong.

Dave also indicated that the Keyword regens for UK failed several times until he discovered that an Oracle table needed to have more space allocated for it. Dave said that this was another indication of lapses in the documentation: the certification 500 training documentation indicated that you needed to ensure enough Unix space was allocated for the keyword regens, but it neglected to mention the Oracle table space needs.

Continuous OPAC had problems for both westlib and eastlib, but each location handled the c-OPAC differently. Because production server and test server are physically located near each other and have IPs in the same domain they could swap IPs between computers, but what had the promise of being a simple transfer had other complications because there were references within the code that had to be altered in order to get the c-OPAC functioning—much more editing was required than had been expected for this process. Eastlib changed the IP’s associated with the Domain names and there is an expected and unavoidable delay in having these changes propagate through the various Domain Name Servers—an added complication occurred in that UK’s domain name (InfoKat) was overlooked. It was suggested that a more robust continuous OPAC process should be investigated even at the expense of using a data backup that is several days older than what we have used in the past.

Mark Paul stated that the time frame for this upgrade worked well and caused minimal disruption in operations. Upgrading on the weekend resulted in only one work day down time—this was a significant improvement. There was general agreement that this was the case. Most of those present indicated the client roll out worked well. Several people indicated that there were a few problem computers, but generally client installation worked well. UK indicated that many of their computers experienced a failure to register a dll file on the majority of their installations, but this was worked around by running install twice and choosing the repair option.

Mark Paul also suggested that the upgrade process would go more smoothly if the old circulation client off-line mode files could be uploaded into the newly upgraded server. It is an added complication to ensure that new clients are installed in multiple circulation locations and instruct everyone on new operations immediately at the start of the upgrade process.

Next Upgrade

Because our current AIX platform will stop being supported by IBM in December, and the Unicode release is already upgrading the Oracle database to Oracle 9i, there may be additional software and hardware complications if there is a platform change too. EISI needs to discuss and investigate what problems might be encountered by moving the Unicode release to AIX 5. Some institutions may have additional complications because AIX 5 requires 64 bit processors and some AIX users are using 32 bit equipment—this is not a problem for the Kentucky consortium. Unicode Solaris needs to be released and these other issues investigated before we can make a decision on a date for the next upgrade.

Service Level Agreement

We will need to be more specific in some of the language for our Service Level Agreement and be more specific when describing responsibilities for response time, etc., in the SLA.

Future Meetings

December 5th 2003
March 5th 2004
June 4th 2004
September 16th 2004
December 3rd 2004

Respectfully Submitted,

Kelly M. Vickery