AR for Mary (reposted from last minutes): Contact Nathan Lynch to ask him to validate the current scripts/tests for interrupt migration in CPU hotplug testing. AR for Martine: 1. Get further info on memtoy test from Lee. 2. Send mail to the SIG to ask for plans/goals for 2006. 3. Send Mary exact kernel version numbers on when hotplug features (CPU and mem-add) made it in. Attendees: Mary Edie Meredith, OSDL Bryce Harrington, OSDL Bruce Vessey, Unisys Martine Silbermann, HP * Report from last DCL TWG F2F in Portland During the last F2F the DCL Technical Working Group decided on a list of top 10 priorities they would concentrate on during 2006. Amongst that list 2 items are directly related to the work done in this SIG. The first one is to do an analysis and write a paper comparing the ballooning approach to the memory hotplug approach for Xen virtualization. The second one is to help complete the technology forecast specifically with regard to hotplug for CPU and memory. * Status on memory remove patches Both ZONE based patches have been submitted by Fujitsu and light versions of the fragmentation avoidance patches by IBM. The feedback seems to be mostly technical so far. * Status on testing (both CPU and memory) There has been a series of hardware issues in the lab that has slowed down the testing effort. Soon OSDL will be getting an administrator to handle the hw issues. Bryce looked at the memtoy tests provided by Lee Schermerhorn but at this point in time he needs more input from Lee to understand if those tests could be easily extended to non NUMA systems. AR for Martine: Get further info on memtoy test from Lee. Bryce is currently working on developing a way to visualize the results from the various tests. * Plans for 2006 - OSDL just received a new x86 box which will be used to run the CPU tests in a fully automated fashion including the report of the results in a visual format. - For memory testing we need to take the existing test plan and turn it into a set of test cases with adequate implementation details. - Proceed w/ regression testing and some performance testing for memory hot-add. - Modify memtoy for non NUMA system and develop other new tests for memory hotplug. - Write a paper comparing ballooning to hotplug memory approaches for virtualization. - Other arguments in favor of memory hot-remove could be * support for node hotplug, * the fact that hot-remove would work both on the iron and in a VM environment, * the fact that the ballooning driver would not be able to handle Windows as a guest on Xen. Those ideas need to be discussed some more. AR for Martine: Send mail to the SIG to ask for plans/goals for 2006. AR for Martine: Send Mary exact kernel version numbers on when hotplug features (CPU and mem-add) made it in. Due to the holidays and various conflicts in people's schedule the next meeting will be exceptionally scheduled on an off-week on January 10th at 11:00am -12:00pm PST, 2:00pm - 3:00pm EST. Happy Holidays to all of you and your families. Martine J. Silbermann
