Microsoft Australia has begun negotiating with TechEd attendees to reschedule technical labs delayed by an infrastructure failure earlier this week.
The experience, caused by a scaling issue of virtual machines at one of the on-site infrastructure clusters, left many of the developers attending the pre-conference technical sessions unable to complete some of their hands-on sessions.
TechEd attendee Hayden McManus said many attendees had been served with "endless loading/progress bars, VMs that wouldn't start, and generally very slow or poor performance".
"Soon, floods of hi-vis wearing engineers were seen clustered around the rack in the corner, worriedly scratching their heads and conferring in hushed tones," he wrote on the Au Tech.Heads blog.
"At the conclusion of afternoon tea, a Microsoft staffer and an engineer from HP were ushered into our lab, and made the announcement that disk queue's were the fault, and that possibly there was an issue with the SAN."
It is believed that only some of the labs were affected on Monday, though a spokesperson for Microsoft did not clarify which ones at time of writing.
Those who attended the pre-conference labs paid $715 on top of the registration fee for the general conference.
The machines had been served by one of several on-site infrastructure clusters established by Microsoft for the conference.
The HP private cloud included 96 gigabytes of memory and 20 terabytes of storage in a RAID 6 form factor. Each virtual machine was established to scale in memory from 512 megabytes or more each.
The SAN failure appeared to be fixed the next morning, with specialist HP engineers effectively re-imaging many of the affected virtual machines to fix the problems.
Microsoft IT professional evangelist and Tech.Ed content manager Jeff Alexander said the core aim of support personnel at the conference was to respond quickly to the issue.
"I've been at pretty much every TechEd and every year there's almost always something that happens," he said.
"We're moving a whole bunch of kit that's generally not meant to be moved; it sits in data centres most of the time and doesn't go anywhere.
"I demo[nstrate technology] all the time and every now and again the demo gods just fail you for no reason. A lot of these guys realise that."
Though Microsoft had confirmed no overarching compensation for developers affected, they would likely be offered a rescheduled time slot to complete the chosen lab.
"We really take the feedback personally, we take it to heart and we want to make sure they get a good experience so we want to give them the opportunity to redo this later during the week," Alexander said.
Attendees reach wi-fi records
The opening keynote of the annual conference on Tuesday night saw a total 896 attendees logged onto the conference's wireless network at an average lag time of a millisecond each.
It represented more than 90 percent of the 985 attendees at the keynote.
The 2011 conference marked the second consecutive year Microsoft had used a dual-stack IPv6 network for internet connectivity to attendees, following last year's successful trial of the protocol.
The core network backbone of the conference featured a direct gigabit link from Tech.Ed's Gold Coast location to a Brisbane exchange, supplied by Telstra.
Organisers had also established a 500 Mbps fallback link from Queensland data centre operator Over the Wire but that, along with other fallback plans from individual conference speakers and exhibitors, had yet to be used.
Correction: The original copy for this story incorrectly asserted that the SAN failure impacted Microsoft's Tech.Ed exams rather than labs. iTnews apologises for this error.
James Hutchinson travelled to the TechEd 2011 conference as a guest of Microsoft.