= Attendees =
Marcos Matsunaga (Oracle)
Aravind Gopalakrishnan (AMD)
Lars Kurth (Chair)
Ian Jackson (Community Rep)
= Agenda =
* Windows Licensing Update (Lars)
* APM ARM Boxes (Ian)
* Test Lab Update (Ian)
* OpenStack CI Loop Update (Lars)
== Windows Licensing ==
Ian: Do we need to implement the capability to set a lower limit than one Windows VM per host
Lars: I don't believe so for now. Even if we doubled our capacity to 48 x86 hosts, the cost per year would still be less than $4800
Lars: However, we may need this capability for next year
Ian: Agrees
Marcus: Can the 24 VMs be run on any machine (e.g. Windows instances on one host)
Lars: Correct, from a legal perspective the only restriction would be the number of VMs that are run at any given time in parallel
Ian: points out that in the past we did not have enough capacity to do this, but we may want to do this in future
*There was also some discussion on how MS enforces license compliance*
Lars believes that MS occasionally requests audits where we have to provide logs on what we ran. But this will need to be confirmed
== APM ARM Boxes ==
IanJ and IanC agrees that taking the donation is a good idea
We do have enough rack-space to go forward
Obviously details on how to connect, etc. need to be worked out
Marcus: asked whether the machines are in our power budget?
Ian: believes so, but will check
*WG agrees that we should recommend to the advisory board, to take the X-C1 EVK kits into the Xen Project Test Lab*
ACTION (Ian): check assumptions (power, connectors, â)
ACTION (Lars): bring up at Tuesday Advisory Board board meeting and get a decision on whether to accept APM's offer
== Test Lab Update ==
18 of 24 machines are now in service
The C600 storage controller issue has been resolved (aka we have a working work-around)
Of the remaining 6 machines,
* 2 going to be in service soon: passed 1st set of commissioning tests; waiting for 2nd set
* 2 more pairs of machines that are still not in service because of serial cable issues - hopefully should be fixed soon
* 2 further DELL machines affected by a BIOS bug - ticket open with DELL (Ticket SR# 909346000). Does not look as if it may be fixed soon
We also had an issue with backlog times, which slowed down the push gate significantly. The result was that many commits in April were delayed. This has now been fixed. The issues were
* We had some software bugs which were exacerbated by new Heisenbugs which were exposed because we are testing on new HW
* This led to the pushgate back-log building up
* We fixed the SW bugs
* As a result, the backlog problem has been solved : average pushtime is back to around 1 day (from 3-4 days)
== OpenStack CI Loop Update ==
* Yesterday CI loop became voting on OpenStack Nova (see picture XenVoting.png)
* Planning to write a blog post to be published Monday or Tuesday during OpenStack summit based on the OpenStack decision
