skuhn

35 ms·

331.
by skuhn 11y ago
Every technical problem is boring and monotonous to somebody, the trick is to find someone who can get interested in it and pay them to deal with it. With any luck I'll be paying Foxpass to solve this one for me in the near future.
332.
by skuhn 11y ago
There are superior aspects to the design of "enterprise" hardware (or rather, rack mounted server hardware) that have been mentioned by others: ECC RAM, redundant PSUs and HDDs, motherboard designs that (hopefully) are laid out fo
333.
by skuhn 11y ago
The HDMI dongle is a good solution for most users. It can also be worked around in software (although it is not trivial). It's kind of annoying that OS X requires jumping through these hoops, but at least there's a few solutions o
334.
by skuhn 11y ago
It is a complete no-no, even if your DC doesn't say anything to you about it. Just think of the potential loss of equipment if the sprinklers were to activate in a datacenter suite -- and if the flammable material is located in your ca
335.
by skuhn 11y ago
The refurbished machines are a decent deal and I haven't had any problems buying them for personal or business use in the past, but the big limitation is on supply at any given time and the configs offered. Refurbs are largely stock co
336.
by skuhn 11y ago
(I wrote the imgix post you linked to) I think Steve has done a great job on his MacBook Pro rack design, but I do wonder about the advantage the MBP has over the Mac Pro for their workload. It may be that the GPUs of the Mac Pro don't
337.
by skuhn 11y ago
A standard server rack (fully populated, just regular app or web servers) can easily be 400k+. Everything built to run in a datacenter is pretty expensive, even if you've done your negotiating work and aren't foolishly paying list
338.
by skuhn 11y ago
There are ways around the lid opening thing as far as activating the GPU (same issue applies to Minis and Pros that have no monitor attached in a datacenter environment). The big requirement for keeping the lid open is probably to control t
339.
by skuhn 11y ago
You don't need the thunderbolt adapter, but if you want gigabit Ethernet it's either thunderbolt or USB3 (as you mentioned). One of the perks of the thunderbolt adapter is that it doesn't require any third party drivers, whic
340.
by skuhn 11y ago
The imgix sandbox is meant as a tool to experiment and try things out. Behind the scenes it uses imgix's production infrastructure via the published API. We think it's really helpful to be able to quickly iterate over ideas, and w
341.
by skuhn 11y ago
I think that's a mistake. You are using abandoned software: djbdns 1.05 was released in 2001. It even has a published security flaw from 2009, for which the $1000 guarantee was paid by DJB, and yet there is still no official release to
342.
by skuhn 11y ago
Entrust, for instance, allows you to purchase certs essentially on a subscription plan. You may have one valid cert for 3 years (and pay the reduced 3 year rate), and reissue it with 24 hour expirations every day. I personally wouldn't
343.
by skuhn 11y ago
(I work at imgix) To follow up on Kelly's response: Our service is only for use with images provided by our customers, and our terms of use specifically require them to have the associated rights for the images they use with us. From a
344.
by skuhn 11y ago
The problem with putting forward a 'do-one-thing-well' rationale is in considering TLS to be a separate problem from serving HTTP. It simply is not, and even in 2006 the writing was on the wall: HTTPS will be the standard web tran
345.
by skuhn 11y ago
Yeah, it's worrisome that people still use the bump version un-patched. The last commit to the official stud github repo is from 2012: https://github.com/bumptech/stud/commits/master (although there are
346.
by skuhn 11y ago
That's true, and you shouldn't support TLS compression anyway (to resolve attacks like CRIME). You should also set SSL_MODE_RELEASE_BUFFERS to reclaim memory from idle SSL connections.
347.
by skuhn 11y ago
They're using the Amazon Linux AMI (2014.09 most likely). It might be the default Apache config that ships with it, since it's super terrible. The F could be easily fixed with a slightly revised cipher suite list, but it would sti
348.
by skuhn 11y ago
It's pretty telling that 20% of the public IPMI hosts are Supermicro. I suspect that a lot of that is completely unintentional. Supermicro has a completely brain dead default behavior, where if you connect power to the machine and the
349.
by skuhn 11y ago
Cost is one factor, but SNI also has performance and IPv4 exhaustion repercussions as well. An SSL cert chain with 30 SANs is 4.3kb, versus 2 SANs at 2.5kb. This is data that is sent for every new connection, and nothing can happen until it
350.
by skuhn 11y ago
URL shorteners are a great idea, but for very limited use cases. I think they only make sense when run by a particular service for use with that service's URLs, like t.co or db.tt, and live and die in tandem with the underlying service
351.
by skuhn 11y ago
> I wish I had the chance to work on something like this! I should have added, for anyone reading: you do! imgix is hiring, and if you don't see a job description that appeals to you, just reach out and let's see. I'm writ
352.
by skuhn 11y ago
Actually, I realized that we were both wrong on the math. 2200 gflop or 3500 gflop are the specs for just one of the Fire Pro cards. Whoops, I was writing a lot of comments that day. So a Mac Pro with D700 GPUs has 7000 gflop/s and run
353.
by skuhn 11y ago
I wish that I could share some of the internal cost analysis that was a big part of the decision process; I've dropped breadcrumbs here and there, but exposing the whole thing just isn't a possibility. Physically, the Mac Pro itse
354.
by skuhn 11y ago
Thanks, the photography was done by our lead designer, Miguel. I am super impressed at what he's been able to capture in an environment that can easily come off as utilitarian and sterile. He has some other work online that you might e
355.
by skuhn 11y ago
Indeed, it is. Servers are expensive; everything built for a datacenter is expensive. It's very easy to make a passive chassis for far less than conventional datacenter gear prices. A lot of cost estimates have been thrown around (here
356.
by skuhn 11y ago
Thanks, I was aware of AutoDMB and Munki, but the rest are news to me. We'll check them out.
357.
by skuhn 11y ago
I mentioned this in another comment, but my understanding (without having worked there) is that they do not use OS X or their own hardware internally. I think that things can be quite different between orgs though, with some adopting a more
358.
by skuhn 11y ago
I think our solution is pretty similar in concept. We just deploy the base OS (with enough config done to ssh in later) via NetRestore. Then whatever packages or setup tasks are required is done in a post-install step using Ansible.
359.
by skuhn 11y ago
Thanks, the article is hosted by Exposure (one of our awesome customers), so I'll pass this note along to them.
360.
by skuhn 11y ago
Not certain if I understand your question, but I'll take a shot at answering: I expect a useful life span for any datacenter equipment of 3 years. A Mac Pros list price is about $4000. We pay less but I'll use public figures throu

More ›