Time and Location of Meeting
April 2, 20143:00 pm - 4:30 pm Main Library 428
Agenda not yet available.
Bill Mischo, Tom Teper, Lisa Hinchliffe, Michael Norman, Jim Dohle, Lynne Wiley, Sue Searing, Robert Slater
Approved as submitted.
Proxy Outage – What happened? Why? Plans to address in future?
It was limited to EZproxy off-campus outage. Lasted from 6:30pm on the 21st to 12:30pm on the 22nd. Approximately 500 users (bots make it hard to state for certain) encountered access issues, and one ticket from faculty/staff (Bill Mischo) on Friday the 21st at 11:44. 11:55 Stephanie baker called help desk> AITS > Jason Strutz. As 12:33 22nd IT started working on fixing the problem. By 12:30 PM the 22nd, the problem was resolved, for a total outage of about 18 hours.
Post incident analysis revealed there was an old/bad snapshot of the virtual machine that grew and consumed the entirety of the virtual disk for the server, locking up the server.
Initial response was slow due to staffing levels in library IT over that weekend, as well as use of/routing of early emails reports/OTRS tickets (normal operating procedure during library IT business hours) versus the after-hours reporting mechanism of calling in the problem, which routes the emergency to AITS and from there to relevant individuals in library IT.
Library IT will first be addressing the underlying issue, followed by an investigation how to modify our automated system monitoring service to send alerts to Library IT when systems enter hung but not completely failed states.
The other reporting piece that was missing was communicating the problem out to our patrons. We have systems for doing that (e.g. gateway messages and system status page updates), but we need a checklist/procedure and standards message templates for rapid response use in these situations. Library IT recently drafted (and had approved by AC).
Progress with SFX Migration to Hosted Solution
Library IT spoke with Emory earlier this week, about their experience with moving from locally to a hosted solution. Library IT has a teleconference with Ex Libris SFX support next week to follow up with some possible issues with the type of Ex Libris hosting we choose and how that effects timing of updates (our choice versus Ex Libris), system access (root or not), and how we can retain the current hostname (http://openurl.library.uiuc.edu or http://openurl.library.illinois.edu). Library IT will have a full report for CAPT after next week’s teleconference.
Web scale team update
April 9th meeting with John Wilkin and the AULs to bring them up to speed on work on the project so far, and near term options going forward. In particular, the capacity issue is paramount (we can’t add new collections now, and within the next few months growth of current collections will exceed our remaining disk space). They can moved us to a cloud-based multi-tenant setup for $19,000 additional per year, or move us to a dual-server setup for $40,000/year, both of which would put us over the bid limit. John Wilkin and John Straw had a meeting recently about this, after which John Straw contacted his admins and the response was we would not be allocated any additional disk space without paying more for it.
Discovery and Delivery Study Team
Given where we want to go long term, what implications are there for that in the short term? Ex Libris has essentially backed us into a corner that will require us going out for RFP, and if we pursue that route, there is a good chance Primo would not be the commercial system (if the group recommends any commercial webscale discovery system) that was selected through an RFP. Given how much customization and direct control we have over Easy Search, it’s possible the group would recommend not implementing any other webscale discovery system, and instead expanding support and investment in EZproxy. The group will be putting together a “pathways” document that outlines all the reasonable ways to move forward with how we provide this type of federated materials access.
There should be a position description fleshed out in case we decide to move away from commercial web scale discovery systems, and expand the capabilities of Easy Search.
The DDST team also has some items they want to recommend to CAPT. DDST has discussed the issue of having three separate catalogs linked from the gateway, and they would recommend the Classic Catalog link from the gateway (not other pages, just from the gateway). Although CAPT is the group that would make the decision to remove the link, they aren’t certain who should be tasked with communicating this change is going to happen. It should include public service and staff training folks. We’ll move this as a recommend action to CAPT for our next meeting, float the idea of removing the classic catalog link to AC, and suggest the membership for the group that will communicate this change to the library faculty and staff, as well as (if any) communication out to the public.
Brief update on Repository planning team work
This request went to EC. A number of people on EC (and not) wrote responses. Key CAPT members are absent today, so we will continue discussion of this at the next meeting.
Meeting adjourned at 4:15 PM