cagataygurturk

Bending Spoons is where products go to die.

What is your hypervisor technology? How do you ensure isolation between tenants?

PS: I see that your VMs are created in CoreWeave datacenters. I wonder the relationship between your company and CoreWeave. Are you just a reseller?

A slight correction: FDP is the fourth largest power in Bundestag with slightly more than 10% votes it got in the last federal elections, and what is happening here is not called taking the goverment hostage but a governing party excercising its influence using the instruments given by the parliamentary democracy.

I feel your friend little bit exaggerated the situation or simply had bad luck.

The staff of the doctor you usually go knows you and don‘t ask your statuary health insurance card upon each visit. They have it in your file.

Worst case you present a paper you download from your insurance provider‘s website.

As a last resort, you declare you are privately insured and you walk out. They send the invoice to home and you send it to your insurer or pay from your pocket. A simple GP visit does not cost more than 40-50 Euros.

People are helpful. Maybe the staff of their doctor was not that day.

Here in Germany, "Mandibular advancement splint"[1] is a popular device recommended by doctors for less severe apnea cases. I was diagnosed with a light apnea and now I am getting this device as my test results showed that a CPAP treatment would be overkill but this could help. I am afraid I can't share more experiences as I am waiting the device to be prepared after 2 hours of dentist visit where they took some measurements.

[1] https://en.wikipedia.org/wiki/Mandibular_advancement_splint

The ARM nodes which are exactly the topic of this submission could be a place where you can self-host your GH runners :)

https://github.com/actions-runner-controller/actions-runner-... works quite well with its autoscaling. You could create a zonal GKE cluster, which is in free tier, and create a small spot vm node pool with ARM nodes. It wouldn't be entirely free but it would cost quite low amount of money.

You can already self-host GH Runners on ARM.

What is your product‘s difference with Cloud Run?

> I can’t think of any of the big tech companies that even use VMware.

You are quite wrong…

I have 3-node vSphere cluster with 36 cores, 300 GB of RAM, bunch of different disk options (magnetic, SSD, NVMe)

It is all powered by Ubiquiti networking gear.

VMware VUG is providing free and legal VMware licenses. I get to run VSAN, on top of it I deploy Kubernetes and I can decide if my pod wants to use poor performance VSAN disks (like EBS on AWS) or faster local NVMe‘s.

These are just couple of things I can make at home. Everything started as an experiment and it started to look like a replica of what we can have in a datacenter, but without a SLA.

First of all, all these stuff is extremely fun for me. Secondly, I won‘t go into details but as a cloud consultant, this thing I started as a hobby became one of the big boosters of my career. I never expected such an effect but I am extremely impressed by this.

> <style type="text/css">

> <!--

> a:hover {text-decoration: none; color: #000000}

> -->

> </style>

The <!—- —> wrapping the inline CSS made me chuckle. The first CSS hack in the history.

I am happy to hear that I was not the only nerd interested in enterprise software during middle school years. You know, while my friends were waiting the latest games, I would wait for the next Service Pack of Windows 2000. Constant exploration of AD, firewalls, networking, purely for curiosity and fun.

Now those friends ask me how to get an IT job by the way :)

Aren‘t preemptible/spot instances a way of dramatically reducing the public cloud cost if the training jobs are designed to be resumable/resilient to interruptions? Most providers offer also GPUs with this pricing model.

The author is looking for Google Cloud‘s IAM architecture.

I was so sick of creating and deleting temporary Route53 Health Checks to measure the latency of my API endpoints from different parts of the world, I ended up creating a CLI tool doing that for me

https://github.com/cagataygurturk/global-load-test

To pull the centre of weight to the bottom, typically overboard space is used only for empty containers in cargo ships. Otherwise it is very very challenging to keep the vessel stable. Actually mostly impossible.

You might ask why they even carry empty containers, it is because not every loaded container goes back to its source location with other type of goods and cargo ships should always carry empty containers to solve one of the most challenging problems of the industry, which is called „empty container repositioning“. Cargo traffic across different locations of the world is not perfectly balanced. So, for example Chinese should get some empty containers back to load them back and send to Western world :)

Source: I am a Naval Architect by education and worked at cargo ship design and construction for a while.

Probably they lost empty containers.

Musicians https://www.forbes.com/sites/quora/2017/10/03/programmers-an...

This decision is penny wise and pound foolish.

Learn how large-scale open source projects are managed and governed. Set up a team culture that could support remote working, even though teams are at this moment co-located. Can your team keep generating the same output even when they don‘t see each other everyday holding meetings? Such teams tend to leverage more async and written communication, write more complete, comprehensive documentation.

A good code review culture also improves knowledge sharing: We sometimes have two paragraphs of explanation in a pull request that changes one line of code. Those pull requests are forming a part of our institutional knowledge base and they are oftentimes referred during future conversations about past decisions.

The team should understand that spending a bit time today to write things down will save much more time in the future when the same thing has to be explained again and again. Also the team should be encouraged to spend this time on writing documents and not punished because of not starting with the next game-changing product feature.

eBay Q2 2019 GMV is $22.6 billion whereas Shopify's is $13.8 billion.

eBay also owns a classifieds business with around $250 million quarterly revenue which does not originate from GMV due to the nature of classifieds business.

Going to EKS would take less time than exploring hacks.

I am surprised that noone mentioned about the bottle collectors phenomenon in Germany. [1] Basicly you pay a deposit (.25 eur for plastic ones, 0.08 for glass.) for each beverage you buy and then sell those bottles back in supermarkets. This is maybe the easiest and most efficient way to force people to recycle. Are you lazy to do that or you don‘t have a place when you finish your beer walking in the street?[2] Then you simply leave the bottle next to a trash bin and in couple of minutes someone takes it to cash it back and call for a day. It‘s a serious revenue stream for a large group of low income people.

[1] https://www.theguardian.com/world/2014/jul/02/german-cities-...

[2] In a country famous with rules and strictness, it is also surprising to see stuff which are not forbidden or no one simply cares, such as drinking on street or public transport, quite openly shopping and using drugs, smoking in bars etc. However some of these might be true only for Berlin, which became a island of freedom in the middle of Europe.

In Berlin there are still people doing this.

Legacy code.

Care to elaborate how you were affected

I know you wonder, yes in Dutch part there are coffeeshops where people can buy pot.

In Turkish, orange is called „Portakal“. Reason is, this fruit was introduced to Ottoman soil by Portuguese sailors. Probably Turkish people heard the name first from North African Arabs living in Ottoman Empire, and they started to use this word.

Year 2006, we had a very high-traffic website running with 1 MySQL server and 1 web server (PHP). Maybe high availability or resillience terms were not coined yet, that's why we were comfortable with having one server per each function. Web server had two ethernet cards, one is with private IP and one is publicly accessible IP. After a while, the platform started to crash and I would be called by my loyal users before Pingdom alerts reach to me, then I would call the datacenter technicians to press restart button of the web server. Obviously it was a lengthy process for recovery, with a lot of human involved.

After a while, I discovered that the issue was about web server's ethernet card attached to internal network and used to connect MySQL server. When that ethernet card stops working, the platform would crash. On the other hand, it was also possible to connect to MySQL using the other ethernet card via public IP. It would reduce the performance of the platform, since all the bandwidth of that card (100 mbps!) is already eaten by HTTP traffic, but at least it would keep it running.

I ended up writing a script at my home computer, checking if the platform is up or not. Once the faulty ethernet card fails, it would connect to FTP, change PHP configuration to use the other ethernet interface to connect to MySQL server, and send an e-mail to datacenter technicians to press restart button.

This script successfully did its job during 3 months, until I eventually replaced the faulty ethernet card and fixed the issue.

Isn't it "Invent and Simplify" like Jeff Bezos says?