The Foundation sponsored Mark Linimon's travel to California from October 28th to November 9th, 2012, to attend MeetBSD, the preceding Developer's Summit, the postceding Vendor Summit, and to perform work on the FreeBSD systems at ISC. Here is his trip report:
I gave a brief overview of the state of the Ports Collection, and invited those interested in more details to attend my talk at MeetBSD proper.
In addition, I answered a number of question in informal sessions from various developers and contributors, some of whom I had met before, and some whom I had not.
Among the issues brought up were ports on various architectures, cross-building ports or perhaps building ports under emulation, and plans for Content Distribution Networks.
Of particular note, I was able to interact in person with the Yahoo team (Peter Wemm, Sean Bruno, and Ben Haga) to talk about issues affecting our machines in the cluster.
I presented a session about the state of the Ports Collection on FreeBSD, and future directions. You may view my slides online.
In particular, I focused on the progress in our new packaging system pkgng; our reworked options system optionsng; the state of ports vs. clang; and the Redports distributed build testing system. It should be noted that hardware for both production package building and also Redports was purchased via Foundation funds.
I also attended other sessions that I found useful, especially Sean Bruno's explanation of qemu. I hope that we will be able to use this for cross-building ports in the future.
At the Vendor Summit mostly I was an observer. However, I did present a brief explanation of the status of the ports tree, and how that might affect the vendors. As well, I chatted informally with Tom Hanrahan of Microsoft's Hyper-V effort about how best to interact with the FreeBSD community. In addition, I was able to interact with a number of other representatives from vendors who had questions about the Ports Collection, particularly about building for non-x86 architectures.
The ISC systems needed some major work.
For one day overlapping my trip, Simon Nielsen was also on-site and did the preliminary install of 7 new Atom servers. These were purchased by the Foundation to be administration servers for a future fallover facility for the main Yahoo systems. I continued this work and did other work as follows:
- I moved most of the Ethernet cables from the old, unmanaged, switches to the Juniper EX3300 managed switch that Simon had also installed.
- I wired serial cables to the new Atom servers and verified that they worked properly.
- I returned one of the Atom servers to iXsystems for an RMA. It turned out that the CPU had suffered early failure. I reinstalled the system once they were finished.
- We still had several systems in rack B2, which ISC wishes us to vacate. Since the portmgr-owened systems there were offline due to other reasons, I relocated them. Two systems still remain.
- I deinstalled a power controller that had failed.
- The cable situation had become sub-optimal. In many cases, the only cables that had been on hand were 25' or more. It was barely possible to get to the back of the systems, and airflow was a problem. I installed various shorter lengths of cable, and correctly color-coded them while doing so.
- I verified all the connections to the console server and the PDUs.
- I attempted to install a new hard drive on our PowerPC server there. However, without a CDROM drive, that system is hard to work with. Instead, I carried the hard drive back to Austin, where I will format it up on my similar system here and send it back to ISC.
- I purchased a second power supply for the Coverity machine.
- I removed rack slides from the systems on which those slides do not fit into the shallow racks there.
- I purchased some rack shelves that will stand in for these slides to help us manage our physical space. They still need to be installed.
- I physically labeled all the systems.
- I annotated all the above in our administration database.