ZendCon 2011
17 - 20 Oct 2011
at Santa Clara Convention Center
Talk comments
Overall a great intro to Hadoop. Learned a ton and gave me enough info to go back and dig into it a little more.
This was a great (albeit brief) look at how ZF2 is going to support DI in various different ways. There was a ton of material to cover. It would have been good to make this session multi-part or longer to be able to show all the examples or maybe go into some of them a bit deeper. As Ralph said some of this does look a little bit black box at this point, but it's a great start. Looking forward to playing with this stuff on my own.
Non-working, demo, but everything else was great. A lot of excellent information, and new to me.
Sorry, -1 for failed demo :)
Who said that Germans have no sense of humour!
Awesome intro to using Hadoop, and learned a lot despite the technical difficulties during the live demo.
I feel that the corporate Oracle stuff took a lot of time away from the interesting stuff. There also wasn't a lot of technical 'meat' to the talk. I think that might have been due to Oracle not talking much about un-released products, but even to discuss many of the new features in MySQL 5.5 would have been much appreciated (I'm trapped in Debian pre-packaged-binaries 5.1 world and haven't been able to play with 5.5 yet).
The memcached and NoSQL integrations coming to MySQL 5.6 sounds awesome, I will definitely be researching those. Perhaps for the next discussion go into more detail about those?
I would work on presentation just a little bit here. A lot of um's and ah's makes it a little tough. Especially at 9:45AM.
Liz hurt my head with so much info being dumped in, but it has made my understanding of event-driven programming a little more clear and will make examining and evaluating things like Twisted and Node easier.
Learned two new things: the Elvis operator ?: and the use of whitelist mappings to help prevent injection of arbitrary strings.
Very rare to find a DBA who also has an extensive web-based programming background like Bill.
Lots of info to digest in just one hour, I look forward to a little bit more maturity and examples on how to move from ZF1 to ZF2
Thank you all for being so patient during the hiccups in the demo.
The problem was that Hadoop bound itself to the wireless network interface on startup (it has a few bugs in that regard, especially when binding to 0.0.0.0) and the active firewall prevented the local task tracker instances from correctly connecting back to the job tracker to finish the reduction process.
I had to disable the wireless interface (or stop the firewall) and restart all Hadoop processes to get it working again, but that of course cost a few minutes of valuable demo time so I was not able to show off how easy it is to add a second computer to my one-machine "cluster" to pretty much double performance.
Apologies again for not immediately identifying the active Wi-Fi, definitely took me too long to catch the problem. I hope you enjoyed the rest of the talk regardless :)
Slides are now up, too.