SCALE 11x - Day 2 Report
Saturday at SCaLE was fun. I staffed the OpenVZ booth most of the day... from 10 AM until 6 PM. I did get a few breaks and a break for lunch. I had a number of people stop by the booth who mentioned they were using Proxmox VE and really liking it. Quite a few people had no idea what OpenVZ was and so I would start off back asking them if they use any virtualization products and the answer was usually yes. If they were a hobbyist user, their answer was usually VirtualBox. If they were a business type person their answer was usually VMware or KVM. I was ssh'ed into a couple of remote machines that were OpenVZ hosts so I was able to show what containers looked like by doing a pstree on the host and showing multiple init processes.
A guy from IllumOS dropped by the booth to ask if anyone was running KVM virtual machines inside of OpenVZ containers. Not to my knowledge. He said they were using Zones isolate KVM VMs on IllumOS. He seemed to have some concern that KVM wasn't secure/isolated enough and that users might be able to break out... and that the zones would keep everything safe. I mentioned sVirt for KVM but I've not used it myself.
Kir gave his presentation at 3PM on Checkpoint and Restore In Userspace (CRIU) which is a sub-project of OpenVZ. He said it went well and he filled his time but he didn't get a chance to actually show a demo... which was a shame because he had a nice video that showed it from beginning to end. Hopefully I can get him to share that video online RSN.
Kir also mentioned to me the commercial containers product that Parallels has to do Windows containers on Microsoft Windows... and that it was a monumental product with a high level of Windows internals knowledge on display... and that they were trying to work with Microsoft to share information... but that Microsoft didn't seem to interested. They are toying with the idea of possibly releasing OpenVZ for Windows but it seems unlikely. Containers on Windows has to bypass some kernel anti-patching technology Microsoft has so it can install the container functionality and then it has to re-enable it to keep the bad guys out. I'm not much of a Windows person and I definitely don't know Windows internals at all... but it was interesting even when dumbed down for me. :)
For lunch I thought I'd head over to Carls Jr. again but that place was packed... mostly with SCaLE attendees... so I ended up walking about a mile to a Burger King. That was a nice bit of exercise. Speaking of exercise, I decided that I wasn't going to use the elevator and take the stairs... which is a bit of exercise because as you may recall my room is on the 12th floor. I did that about three times down and up on Saturday.
A long time OpenVZ user named John Wenger from the L.A. Co-op stopped by to visit with Kir but he was away from the booth at the time. Late in the afternoon a guy from the Zenoss booth stopped by to say that he was going to give a demo using ZenPack (or something like that) that started up a few OpenVZ containers and showed off the monitoring capabilities of Zenoss. I commented on his "Bring IT" tee-shirt. After I had been talking with him for about 10 minutes it turned out that he was Daniel Robbins... who I have talked to a number of times on the #openvz IRC channel. Perhaps you recognize his name. Daniel was the creator of the Gentoo Linux distribution. These days he works on Funtoo. I jokingly asked him to sign my arm but he said I was weird. :)
Maybe it is because I'm in the L.A. area and I watch a lot of TV but one guy stopped by the booth who looked a lot like Hugh Laurie. If you don't know who that is, look it up. Anyway, he didn't know who Hugh Laurie was and hadn't been told he looked like him before. If only I had taken a picture. Anyway.
A few people reported that Java-based apps (like Jenkins I think) didn't perform well inside of a container and tended to spike I/O usage that would make the machine unresponsive for periods of time. Told them to check bugzilla.openvz.org and jump on an existing bug report if one existed or to file a new bug.
I got a chance to walk around both exhibit rooms and take pictures of all of the booths... but once I got them copied off of the SD card to my netbook I discovered that the vast majority of them were terribly blurry. I guess I was in too big of a hurry and wasn't holding the camera still long enough. The lighting in the place isn't that great but I didn't want to be pulsing a flash in everyone's eyes. Anyway, I dumped that set of pictures... so sorry, no pictures today like I had promised. I will make a concerted effort to take all new pictures.
I went by the OLPC booth and saw Caryl Bigenho. As you may recall, her and her husband Ed have a summer home outside of Bozeman and have visited every year. Caryl gave a presentation to the campus on the OLPC about two or three years ago... and she is the one that suggested we apply for a lending lab which netted us 10 OLPCs to share with the Montana community. I ended up going out to dinner at Denny's Restaurant (that is about half a mile from the Hilton) with Caryl, Ed, and a friend of theirs named Tony. Tony told me some about various OLPC deployments he has helped with in a few different countries in Africa. That was very interesting.
At 9PM I attended a Raspberry Pi Birds-of-a-Feather (Bof) get together. Caryl Bigenho lead it and asked everyone who had one or more Raspberry Pis and what they were doing with them. One guy who I think had an Australian accent but said he visited England at least once a month and that his son was involved with the MagPi magazine Kickstarter project. He said his son was now working on another Kickstarter project for an add-on power button for the RPi... and he showed a prototype microSD card adapter for the RPi that was different in that it was NOT the normal SDcard size. It was specifically made for the RPi SDcard slot so you could put a microSD card in it, plug it into the RPi and it does not stick out from the slot and is flush with the side of the board. I won't go into the various projects that people were doing because it was a very broad list (ham raid, First Robotics projects, co-location for RPi as a hosting server, etc). There were a couple of people from the Tiny-Core Linux booth and they said they had recently gotten Tiny-Core going on it and that it was the smallest and fastest Linux distro yet. They are showing it at their booth for anyone who wants to stop by and see it. Caryl mentioned that she had recently acquired an SDcard with Sugar (the OLPC learning environment) on it for the RPi. She hadn't had a chance yet to try it out but hoped to later in the day on Sunday.
There is a huge amount of interest in the RPi and even though other competitors may come and go in the space, the RPi has the numbers and add-ons and the people excited about it... that it should remain a viable platform for at least a few years.