RyanGWU82

That's not actually how A records work. Clients are supposed to randomly select one of the A records, although this doesn't always distribute the load as randomly as it should. (My company's site used to have 4 A records for 4 load balancers, and we found that one of them received 25-30% more traffic than the rest.) Clients certainly don't have a mechanism to retry if one of them fails -- they'll be stuck with the failing IP for the length of the TTL.

Isn't that the whole point of the story? Zuck paid income tax on the initial grant's value (which may be par value), and the stock's appreciation will only be taxed at the reduced capital gains rate?

What concerns you about S3?

That's not what I've seen, at least for FileVault 2 under OS X Lion. I'm using a 2010 iMac with a third-party SSD, and a 2011 MacBook Air with Apple's stock SSD. Both computers are blazing fast at disk access, even with full-disk encryption enabled.

If you're confident you're encrypting all your data, then you're still way above average. The nice thing about FDE is that I can "set it and forget it." I don't have to think about which files belong on which partition anymore. It's all safe.

Nah, this has been in the works for a few months. I deleted 160 million objects in the spring, which was enough for me to get a phone call from an Amazon product manager. He asked if there was anything they could do to help the deletes go smoothly, and outlined this feature way back then.

We've got 160 instances affected, which is obviously going to require a pretty massive amount of work. I'd estimate we'll be spending about 30 hours on this project over the next week. Yes it sucks, but asking you to reboot your servers once a year isn't exactly a huge hardship.

It's easy to forget that the cloud is powered by real physical infrastructure, and that cloud vendors have maintenance needs too. AWS gave us a fair amount of notice about the reboots, and they've been understanding and helpful along the way. I'm willing to be understanding toward them too.

I brought StartSSL to my last 2 employers for internal sites, like development sites and internal apps, and it's been wonderful. We haven't found any problems with StartSSL certificates in browsers, from IE6 up through the newest browsers. We've considered using StartSSL for our public sites too, but haven't done so yet.

Do customers really care about the name brand or "warranty" of the SSL certificate? What percentage of customers are even aware of who issued your certificate, and do those customers really have concerns about StartCom?

This sounds like the argument we heard from DynDNS salespeople last year, about how our customers expect us to be using "enterprise" vendors, rather than "consumer-grade" vendors like Amazon Web Services. Even our savviest customers have never asked us who our DNS provider was.

This isn't an article, it's a press release.

See, I still don't understand how this is dishonest. They didn't try to make "undeletable cookies," they're not using cookies! I understand that to a layperson "cookie" means "anything that identifies me to a web site," but they're wrong, and what's dishonest about using an analytics mechanism that's not a cookie?

Well, they said that they "made the following changes" -- one of which is not using eTags. So I guess they stopped that.

(That said, they never aggregated this data across multiple websites, so I really don't get what the whole fuss was about.)

It drives my boss crazy that the universal icon for Save is a picture of a floppy disk. We haven't been saving documents to floppy since 1993! But it's also scary to be the first one to buck the trend, so our product uses the same floppy disk as everyone else...

Are you using software like Nginx or haproxy, or a hardware load balancer? If you're using software, then I recommend using something like Heartbeat to share an IP address between the two. It basically does the same thing that mahadri suggested, but it's automatic. Here's how it works.

Let's say you have two servers, primary and secondary -- primary is the one that's currently acting as your load balancer. When both servers are running Heartbeat, they exchange UDP pings every second. If the primary server stops responding, then the Heartbeat process on the backup server will automatically take over the shared IP. When the primary box returns, it won't cause any problems or steal back the IP address, until you explicitly instruct it to do so. (If you really wanted it to steal back the IP automatically, I'm sure you could configure it that way.) You can also initiate a failover with a single command, which is useful when you need to do maintenance on the primary box. Running "/usr/share/heartbeat/hb_takeover all" is a lot simpler than having to manually reassign the IP address.

My company's infrastructure is very dependent on heartbeat -- we use it to support our load balancers and all our back-end services. You need to be careful or it will do "surprising" things when you're rebooting or changing the network config, but on the whole it makes it cheaper and easier to provide highly-available services.

Of course, if you're using a hardware load balancer, it should have a similar failover protocol built-in. I'd recommend just using its built-in features, that's what you're paying the big bucks for.

My company posted 3 job listings 3 weeks ago -- you can look at my posting history to see the jobs I posted here on HN. So far we've gotten about 120 responses in total. The majority were for the Software Engineer and VP/Engineering positions; very few responses for the Operations Engineer job. Most of the software and ops candidates were so-so, though the VP candidates were quite a bit higher caliber.

Most of our candidates came from Craigslist; we advertised in 3 cities and got quite a few candidates from all 3 cities. We also got a few good candidates from the monthly HN thread. We posted to a few inexpensive job boards like Startuply and Startups Hiring but we didn't get many responses. We haven't tried any of the more expensive job boards like GitHub and Stack Overflow because the ROI isn't very clear. Those sites are pretty expensive, especially with 3 positions open. I'd love to hear more from others about whether those are actually worth the money.

Heh, I did the same thing after Startup School 2005. I was in DC but Startup School got me to apply to Stanford for the master's program in CS. Six years later and I've got my masters and I'm having a great time at a Bay Area startup!

The query "is:unread -in:inbox" will find all the unread email that's filtered out of your inbox. It might not answer this question though -- Gmail may still report the total as "hundreds" or "thousands".

I've had a Nightingale CXO for about 5 years and I love it. I went to the store intending to buy an Aeron, but I'm glad I considered some other models that weren't quite as trendy. Wish I had one at work too!

San Mateo, CA

PBworks is hiring a Senior Operations Engineer!

PBworks is hiring a Senior Operations Engineer who can wear many hats. We're looking for folks who aren't scared of being hands on with everything. Willing to compile a grsecurity Linux kernel? Troubleshoot network packet loss? Scale databases? If you’re able to take on a wide range of responsibility, this position is for you.

Your job will be to create an environment where 4 million users and over 1.5 million workspaces can flourish and change the way people work. You'll get the freedom to do the job right, and a fantastic engineering team to back you up. What are you waiting for? Mail your resume and a short paragraph of why you'd like to join to jobs-ops@pbworks.com.

What You'll Be Doing:

- Physical server management: Managing a cluster of about 70 physical servers in San Francisco. Plus 3 servers in Sacramento. Plus a handful in our office in San Mateo.

- Engineering: Developing ops infrastructure tools in PHP, Perl, Ruby, or shell scripts.

- Automation: automating deployment with puppet.

- Monitoring: using Ganglia and monit to understand what's going on with the systems.

- Networking: Running the network infrastructure, including routers, switches, and site-to-site VPNs between the three locations above.

- Databases: Managing a handful of MySQL databases: advising developers on schema changes, keeping the databases reliable, and optimizing performance.

- Security: Analyzing the systems' security, keeping up with new threats, and recommending and implementing new security technologies and policies.

- Email: Ensuring the best possible outgoing email deliverability, and that we don't lose any incoming email.

- Office IT: running the office network; operating the employee VPN; occasionally fixing a printer or troubleshooting someone's laptop.

- Planning: Monitoring resource utilization trends; budgeting; planning for growth.

The Operations Engineer role is extraordinarily wide. If you've got software development experience, there are a variety of customer-facing infrastructure projects that may interest you. If you like NoSQL, there are definitely ways we can add NoSQL databases to optimize performance and developer productivity. We're looking for someone with a wide range of experience and interests, but if you haven't already worked in all these roles, we're here to help you learn.

Requirements: 3+ years experience running a production cluster. Debian and SaaS experience strongly preferred.

Bonus: Knowledge of PHP, puppet, git, quality programming abilities, computer science chops, and an understanding of SaaS business.

Technology: You'll get to solve real business problems with a wide range of technology. Here's a sample of what we work with: Debian Linux, Apache, Nginx, Percona Server (MySQL), memcached, MogileFS, XtraBackup, LDAP, puppet, Ganglia, monit, Perl, Node.js, ejabberd, Vyatta, Amazon Web Services, Solr, Postfix, git.

Why we're different: PBworks blurs the lines between operations and development. You'll sit in on every team engineering meeting and have input to decisions at all levels. As part of your job, you'll continue to architect our system to handle failures so you won't get paged when a hard drive fries. We're committed to providing the resources so that these failures don't become emergencies. We value automation and addressing the root cause of issues so they don't happen again.

To apply, email jobs-ops+hn@pbworks.com.

San Mateo, CA

PBworks is hiring a Web Software Engineer!

We're counting on you to make a big difference in our product. It's not just an opportunity to make an impact at PBworks, it's a responsibility. Millions of people will be using your code.

You're a jack of many trades: you know your way around a Linux command line and Firebug with equal aplomb; Javascript optimizations and SQL optimizations (and when you shouldn't yet bother with them); the design of MVC frameworks and visual design with CSS. You know how to make something beautiful and approachable soup-to-nuts. You've coded a few projects of your own, maybe some of them have even made the front page of Hacker News, but you're curious to jump into an existing small team that actually has millions of users and do some real, helpful work.

Fluency in: HTML5, JavaScript, CSS3, JSON, jQuery, PHP, MySQL, git

Bonus for: node.js, Erlang, Perl, bash, Photoshop, Smarty, Prototype, mogilefs, memcached

You love: Agile, unit testing, functional testing, rapid prototyping.

To apply, email jobs-fwe+hn@pbworks.com.

San Mateo, CA

PBworks is hiring a VP of Engineering!

PBworks is looking to hire a world-class VP of Engineering to direct, manage, and grow our engineering team at PBworks and join a great management team. Candidates should have knowledge and experience with Software-as-a-Service best practices and methodologies, such as frequent deploys, functional testing, and automated rollbacks. Candidates should have at least five years of management experience but still be active programmers with code-review and prototyping level strengths in a modern scripting language (such as Ruby, Python, or PHP) as well as fluency in modern HTML, CSS, and Javascript. Strong preference will be given for LAMP stack proficiency. Extra consideration will be given to candidates with experience hosting software for very large clients and candidates with UX skills.

Candidate should have experience building and executing effective agile product development and project management strategies, and continuing to build a top-notch, highly, innovative team. Experience with workspaces and team collaboration software in general and PBworks in particular is a strong plus. Master's in Computer Science preferred.

You'll be in charge of one of the most powerful, scalable, popular, and sophisticated hosted collaboration tools in existence: with over 4m users a month, hundreds of terabytes of user data, billions of recorded actions, and a million and a half groups. Our team is small but highly effective; every engineer in the group is capable of working on a wide diversity of projects - from low level systems debugging to SQL optimization, to best practices in modern Javascript and CSS. You'll be redefining the future of how our customers get their work done - everything from collaborative medicine to treat cancer patients to innovative creative teams shaping the next great TV commercial or videogame to university research labs, and even a few un-stodgy law firms.

We'd love to have you come help our team win and grow!

To apply, email jobs-vpeng+hn@pbworks.com.

Yes yes yes! This is what drives me CRAZY about the job postings from Y Combinator companies here on Hacker News. I would NEVER apply for a job without knowing what the company does and who is behind it. "Stealth mode" is a lame excuse. We've all learned that execution trumps ideas alone, and is a surprise launch really the cornerstone of your execution?

Before applying for a job, I would want to check out the company's home page and see "do they look competent? Do they look like they're building a product that they're proud of? Is it something that I want to be a part of?" I'd also want to check out the company on LinkedIn or other social networks, to see if they're actually thought leaders in their community, and to see if I know anyone who can give some more color on the people involved.

It's a job seeker's market out there. You may be special, but if you don't tell me who you are, I'm unable to determine whether you're really special or not.

It's cheap and easy to use a distributed network of DNS servers with locations around the world. There's much less network latency when the edge locations are closer to the end user. There are dozens of good DNS providers out there. We currently use DNS Made Easy, and we're thinking about using Amazon Route 53 as well.

If you have multiple web servers servers in different locations, you might be thinking of "Geo-DNS" -- this actually directs the end user to the nearest web server. There are pros and cons to this approach, namely that the end user may be using an intermediary DNS server that's not actually very close to the user. For example, you might end up sending all Comcast users to your east coast USA server, no matter where in the USA they actually are. I'm not familiar with providers who offer this service, but here's a Server Fault thread on the matter: http://serverfault.com/questions/30567/geo-dns-providers

I think you got it backwards. The trick works when you're ending your trip in a hub -- you take the first leg, but skip the connection.

So, for example, say you're trying to fly to Chicago (ORD). The article's point is that it may be cheaper to buy an EWR-ORD-STL ticket, because there's actually more price competition when flying to St. Louis.

And "apt-get install puppet" if you're on Debian or Ubuntu.

At PBworks, when we had about 10 people, we added the position "Head of Operations & IT." 97% of that job is handling operations for our production cluster, keeping it up and running 24/7. The other 3% is typical internal office IT stuff -- networking, email, security, "the printer doesn't work," etc. The tasks aren't all that related, but at a small company, everyone wears many hats.

By that logic, doesn't a hacker lack the skills that designers and only designers possess?

The whole dichotomy is silly. A tech business needs to figure out what to build, to build it, and to sell it. (Not strictly in that order.) Every one of those steps requires cross-functional thinking.

You can get free standard SSL certificates from StartSSL - https://www.startssl.com

They also offer extended validation and other paid options. I use them for wildcard certificates, which are normally at least $200 each. I've created nine wildcard certs since January 1, for a total cost of $50.

If you sit at a desk all day, a good chair is absolutely worth the money. Desk workers are prone to get back problems from sitting all day long. Programmers more so, because we try to get "in the zone" -- and once we're there, we don't get up or move around much.

A high-end chair will probably last for ten years, so you're talking maybe $100 per year. Compare that to the cost of regular massage, or chiropractic, or back surgery, and it's clearly a good return-on-investment.

But the Aeron isn't the only good chair around. I learned this back in 2006 when I was planning on buying an Aeron too. I stopped by a brick-and-mortar HealthyBack store (http://www.healthyback.com) to try out the Aeron, but I tried out the competition too. I wound up buying a Nightingale CXO chair instead. It wasn't as trendy, but it fit my body frame a lot better than the ones people usually mention online. It's important to do this comparison shopping in person, because our bodies are each very different and you need a chair that feels right for you.

Stanford doesn't even have those "registrar-imposed limits." Classes almost never fill up; they just get larger rooms or add TAs. Students don't usually register for classes until after the term has started!

I was an undergrad at a more traditional school where we registered for classes months in advance. I was amazed that the Stanford system was so informal, but in practice it worked just fine. It's wonderful to be able to audit a few classes before deciding on exactly what you're going to take; you may learn that you can't stand a professor, or love a surprising subject.

I think you've got the math backwards -- the native capacity is 38.4 TB per unit. If your data is compressable, you can store roughly 2X the amount of data in the same physical space, which is how they come up with the 76.8 TB figure.

There are certainly VCs that would take on large entrenched companies. VCs invested $33 million in Cuil, for example, even though Google had cornered the search engine market long before. And VCs fund companies that require huge capital investment, like medical devices and alternative energy. But both of these issues are big red flags that will make it a lot more difficult to get your company funded.

In a situation like this, I'd expect you'd need to convince an investor that (1) you have a sustainable competitive advantage against the big players, so they can't just take your idea and destroy your business, (2) your business model can generate enough profit to justify the outlandish capital requirement, and (3) you have the right team to make all that happen. #3 is a lot easier when you don't have such big hurdles with #1 and #2.