I have my own webserver on a micro Amazon instance. It seems it's dying under hackernews pressure :/
Any suggestion on something I could do? It's a wordpress blog on apache.
I have chaching enable. It's apache that is dying.
Does anyone know how to transform a t1.micro instance in something bigger or how to throw more resources?
What is your worker configuration for Apache? You can make it much hardier with some tweaking, or you can get even more robust by switching to nginx, or you can hit lightspeed by putting Varnish in front of it (listed in ascending order of complexity)
You could host your blog on Amazon S3. Unless there's a really compelling reason why you need to dynamically generate your blog on every request, serve it up as a static website. That way you can use your micro instance to do the generation and push it up to S3 which frees it up to do whatever other tasks you want while S3 does the serving.
I am a masters student in an developing economy, who will soon join a PhD program in US, and I will do that after rejecting an analyst position at major banking firm.
For me, PhD makes sense because it would give me 4-5 tension free years of research with my PI worrying about the grants while I worry about science. Sure, job prospects might not be great for saturated market in US and Europe, but I hope I will be able to get a academia job back home or some place nearby.
I think system needs an overhaul to create jobs, and first step should be to decouple teaching from research. As there are only so many teaching jobs a university can provide, why not create independent labs? Max Plank institutes seem a fine model to me.
Then its not an university any more is it? You could always go and start an independent R&D company.
Fucking around with uni's to employ people that nobody wants to employ is just fucked up.
Correct me if I'm wrong - but with PhD's its the usual affair - softer the field, bigger the unemployment - meaning that there aren't nearly as many jobless PhD's in physics, medicine and chemistry as there are in economics, literature,...
So what really should be done is maybe stop educating so many people that nobody wants to hire?
I can only tell about my field, which is Physics, and a hard science field.
No, it is still increasingly hard to find a job (in US) from what I saw. Even getting to tenure track is achievement in itself. For example, any major top school (like Caltech, MIT, etc) train as many as one and half to two times the PhD students each year compared to department strength. MIT Physics, for example, intakes roughly ~25 PhD students/year for faculty strength of ~80. That's almost 125 PhD graduating every five year. There is no way all of them would find an academia job. Doing two postdocs is increasingly normal, and I know people doing three-four of them.
So yeah, there are surely lot of good people who are sucked into this broken system, and we sure need some fixes to it.
> Then its not an university any more is it?
> You could always go and start an independent R&D company.
a) MPIs are publicly funded.
b) If you want to do research, do not ever start a company. In terms of overhead, that's even worse than being a tenured professor.
> Fucking around with uni's to employ people that nobody
> wants to employ is just fucked up.
Don't believe that the current system is a product of free market forces. It's bullshit, fueled by weird anachronistic traditions and byzantine funding channels.
> Correct me if I'm wrong - but with PhD's its the usual
> affair - softer the field, bigger the unemployment -
> meaning that there aren't nearly as many jobless PhD's
> in physics, medicine and chemistry as there are in
> economics, literature,...
To a certain extent, that's true. Don't be a non-genius, not well-connected sociology/literature/whatever PhD. You'll be unemployed, or earn 5k$ a year. However, you're severely oversimplifying the actual dynamics.
There's a employability sweet spot on the hard-to-soft-science scale. Computer science, certain (NOT ALL) branches of mathematics and physics, chemistry, engineering? Go for it. Here, the industry-academia membrane is fairly permeable, funding isn't too much of a problem (because large chunks of the industry live off your stuff), and working conditions are good. You can do plenty of consulting on the side.
Theoretical physics? Obscure branches of mathematics? Astrophysics? You're very much fucked. Not as much as an art history major, no, but it's pretty bad. Most importantly, the life sciences (which includes everything from microbiology to neuropsychology): it's a nightmarish rat race. If you do something biomedicine-y, you'll get funding. But the number of people you're competing with is ridiculous. Researchers spend years and years as post-docs, in giant labs, without a chance to get to the top. Some disciplines are highly relevant for medicine etc. but there aren't that many industry positions. Even if you're in virology, competition is fierce -- because unlike chemical engineering, these fields are highly appealing for millions of biology undergraduates.
All these blog posts come from people in life sciences. I'm hardly surprised by that.
> So what really should be done is maybe stop educating so
> many people that nobody wants to hire?
They are, but aim to solve a very different problem. MPI researchers are free to focus on research without having to do half-assed teaching and supervising. They're still subject to the same market-forces ("up or out", "publish or perish", etc.) as everyone else. If anything, they're even more competitive -- virtually all post-docs would love to do nothing except churning out as many papers as possible, after all.
It'd be a different situation if there weren't just MPIs with people doing research exclusively but also non-research jobs at universities. Then the bottom 80% of all PhDs could earn their money teaching the millions of undergraduates enrolling each year instead of chasing an elusive dream (read: tenured faculty job).
> Sure, some postdocs may realize too late they don’t
> really want to be independent and they would gladly keep
> doing what they are doing for some more time: this is
> what positions in industry are for², and this is what a
> lab tech position is for. No need to invent new names
> for those jobs.
a) Industry positions require and offer very different things when compared to mid-level research positions. Lab techs don't do original research and often earn even less than their junior faculty colleagues. What the authors of the blog posts in question were going for is simply a sustainable middle ground in academic research; something between "up" and "out." Realistically, not everyone will find a tenured professorship. It'd be much more efficient to keep these (often well-trained) researchers in mid-level positions where they can do original research without having to manage a huge lab or being drastically underpaid.
b) Moreover, the industry/lab tech idea completely ignores anything other than life sciences and certain physical sciences. There are industry jobs and well-paid lab techs in microbiology, engineering, some areas of physics, chemistry. Other than that? Not many alternatives to the standard career path, really.
21 comments
[ 3.4 ms ] story [ 53.9 ms ] threada) Throw more resources at it. I am not sure how it is done, but cloud can scale on demand from what I know.
b) If not using already, use some static page serving plugin for Wordpress, like WP Super Cache or other. This should reduce some load.
Install a caching plugin for Wordpress.
Copy the output of Wordpress for that blog post to a static .html file and use mod_rewrite to point requests for the post to that file.
Turn off plugins you don't need.
Install a reverse proxy cache (such as Varnish) in front of Wordpress.
Out of the box, the configs which ship with Apache on e.g. Ubuntu are virtually engineered to fail if your blog ever gets a traffic spike.
For me, PhD makes sense because it would give me 4-5 tension free years of research with my PI worrying about the grants while I worry about science. Sure, job prospects might not be great for saturated market in US and Europe, but I hope I will be able to get a academia job back home or some place nearby.
I think system needs an overhaul to create jobs, and first step should be to decouple teaching from research. As there are only so many teaching jobs a university can provide, why not create independent labs? Max Plank institutes seem a fine model to me.
Fucking around with uni's to employ people that nobody wants to employ is just fucked up.
Correct me if I'm wrong - but with PhD's its the usual affair - softer the field, bigger the unemployment - meaning that there aren't nearly as many jobless PhD's in physics, medicine and chemistry as there are in economics, literature,...
So what really should be done is maybe stop educating so many people that nobody wants to hire?
No, it is still increasingly hard to find a job (in US) from what I saw. Even getting to tenure track is achievement in itself. For example, any major top school (like Caltech, MIT, etc) train as many as one and half to two times the PhD students each year compared to department strength. MIT Physics, for example, intakes roughly ~25 PhD students/year for faculty strength of ~80. That's almost 125 PhD graduating every five year. There is no way all of them would find an academia job. Doing two postdocs is increasingly normal, and I know people doing three-four of them.
So yeah, there are surely lot of good people who are sucked into this broken system, and we sure need some fixes to it.
b) If you want to do research, do not ever start a company. In terms of overhead, that's even worse than being a tenured professor.
Don't believe that the current system is a product of free market forces. It's bullshit, fueled by weird anachronistic traditions and byzantine funding channels. To a certain extent, that's true. Don't be a non-genius, not well-connected sociology/literature/whatever PhD. You'll be unemployed, or earn 5k$ a year. However, you're severely oversimplifying the actual dynamics.There's a employability sweet spot on the hard-to-soft-science scale. Computer science, certain (NOT ALL) branches of mathematics and physics, chemistry, engineering? Go for it. Here, the industry-academia membrane is fairly permeable, funding isn't too much of a problem (because large chunks of the industry live off your stuff), and working conditions are good. You can do plenty of consulting on the side.
Theoretical physics? Obscure branches of mathematics? Astrophysics? You're very much fucked. Not as much as an art history major, no, but it's pretty bad. Most importantly, the life sciences (which includes everything from microbiology to neuropsychology): it's a nightmarish rat race. If you do something biomedicine-y, you'll get funding. But the number of people you're competing with is ridiculous. Researchers spend years and years as post-docs, in giant labs, without a chance to get to the top. Some disciplines are highly relevant for medicine etc. but there aren't that many industry positions. Even if you're in virology, competition is fierce -- because unlike chemical engineering, these fields are highly appealing for millions of biology undergraduates.
All these blog posts come from people in life sciences. I'm hardly surprised by that.
How is this not "fucking around with unis"?It'd be a different situation if there weren't just MPIs with people doing research exclusively but also non-research jobs at universities. Then the bottom 80% of all PhDs could earn their money teaching the millions of undergraduates enrolling each year instead of chasing an elusive dream (read: tenured faculty job).
b) Moreover, the industry/lab tech idea completely ignores anything other than life sciences and certain physical sciences. There are industry jobs and well-paid lab techs in microbiology, engineering, some areas of physics, chemistry. Other than that? Not many alternatives to the standard career path, really.
I hate stupid systems. stop writing concrete widths in CSS
I'm very sorry for the off-topic content but i'm old-enough now and I dont have to give a crap anymore. idknow at google's mail