Perform a write with as many children as possible and hit the power button?

Well the test I’m writing needs to probably be automated at some point so that won’t work lol

I wasn’t even sure if sending a SIGSEGV signal to my process would adequately simulate a crash, but i think it does. There’s still partially written data in my database.

What type of database? I know Oracle has a ton of stuff for those specific situations, from the extreme stories I've heard (major datacenter blowing a power grid) they seem to be fairly reliable.

My company has designed a database from the metal up. So we’re working on durability right now because I pointed out very early on that our recovery process was terrible

It’s basically just a super fast keyvalue store. But they optimized most everything for speed and kind of left everything else to the dust

Why did they write their own database? Unless it's your core product or there are very specific (and good) reasons why you must have this implementation I'd recommend looking into the cost of switching to something like Redis.

Maintaining your own implementation is unlikely to add value to the company and could end up being a huge time suck as you grow.

It’s the product.

It works really well at crushing benchmarks, not so well at some of the ACID guarantees.

Well, I would make a joke about needing another database variant but I just applied to supervise data-driven dog walkers so what do I know...

we have other database variants but they're built on top of this one, so i place any issues this DB has (especially ACIDic ones) super duper high priority.

@kerowo - do you think this is a good idea if i sell it well? or will it just generate eyerolls from my more jaded teammates?

Interesting, I remember you PM'd me the company and I didn't realize at the time it was a database company. I'm probably an idiot.

JMakin, it feels like the company should already know what their market wants out of this database and it’s pretty weird that you have any influence in the priority of features like this. That’s not a shot at you but just there’s no way you can know what the most important thing is to work on.

I also find your role super hard to do well. You’re expected to have some sort of management / authority over these people while also spending half your time as a super junior developer that needs serious mentoring to become productive. It just seems like a really hard balance to strike.

If I were you I’d try to push the job into only one of those areas because I think it’s possible for you to be successful in either role but really hard to do both.

Jj,

They literally have no leadership whatsoever. They just kind of have done what they felt like. It’s a bunch of PHD’s working on a science project basically.

All i did was point out a few ancient issues that seemed related and said “hey, this is important right?” And everyone agreed so now we’re doing that.

Ok... that sounds messed up. I guess it’s a good of a place to get experience as any but I’d definitely be planning on moving on as soon as possible.

yea, i mean it's totally dysfunctional, but i like the company, the people, and my boss. Everyone's super friendly and it's a good environment. No one too difficult except the one guy, but he mostly confines himself in his office.

If I don't get **** canned, which I don't think anyone ever has here, I'd like to stay a while - the pay is good for my area and the investment money will still be flowing for another year or two. I think there's an extremely good chance a large company buys our product outright at some point. The concept is there and I do believe in it, it just needs some work. I'll probably learn a few things.

What I'm most concerned about is spending time here and having it not translating well into my next gig - so I am making a commitment to become a better programmer.

That's a good plan. If you want to be a developer, focus on programming. If you want to be a PM, focus on that. Half assing both will probably not turn out well if/when you leave this company for the next one.

I guess I’m just skeptical that it’s a great product and investment money is going to be flowing with such a glaring product development issue. But I get that there’s some crazy money out there.

Our benchmarks are pretty impressive. We have a few customers. I know we show up in tech blogs from time to time. People do use our software, but they don't use all the features. They mostly just use it as a high speed cache AFAIK.

our CFO and CEO are really savvy business dudes as well. But at some point, this thing is gonna need to go to production, and I think that's where it's stalled.

To top it off our CTO/company founder just started this crazy new product that seems like it's going to be enormously complicated and troublesome to develop.

I probably shouldnt say anything more about the company or I could get in trouble somehow, but I really do believe the product concept is good.

It feels like a few contradictions in there. For example, I don’t see how savy business guys, with at least some customer usage, don’t already know the things that should be worked on much better than someone brand new to the company (let alone the industry).

But anyway I shouldn’t harp on this. And I think you’re already thinking of the concrete actions that come from my concerns (mainly making sure you can use this experience to further your career - whatever you want that to be).

They dont work on the development at all

But aren’t they involved with product direction? Do you mind saying how big the company is?

no, only vaguely. It's small. They pretty much take our benchmarks and sell them to investors, I think. Granted I don't know what I'm talking about and can't speak well to this place because I've only been there a short time - but that's the gist of it I'm pretty sure.

Sometimes they'll get an interested customer from the business/sales end that says "Hey we want it to do this, can you make this happen?" and our development side comes back and says yes/no and then it's done or not done.

Occasionally a potential customer comes to us with an issue with a demo/free version of the software. But that's forwarded directly to my boss from the "sales" guy. The CEO/COO/CFO don't touch the products at all they pretty much just sell them and drum up investment.

I mean they've raised a ****load of money for a startup and I did some background research on them and they are legit, 100%. So I have a lot of confidence in them, but it does seem like the end game is to eventually be bought by facebook or something, in which case my shares would probably be screwed because they owe a lot to investors I think.

Even from just selling the product they should know what the most important stuff is to build. It all seems very bizarre to me. But could be stuff lost in translation or context I’m not getting.

I think the issue is, I'm starting to learn, is our CTO/the cofounder kind of runs the place like a dictatorship. he comes in with some crazy idea, commits 7k lines of code out of nowhere, and it's everyone else's job to fix it for the next year. I mean that basically just happened in the last month I've been there.

The POC tests were sold to somebody, then we had a department wide meeting, but he was really vague about it. Then disappears again for a while. It's just how things work there.

I know for a fact 2 people are super unhappy with the way things are but I don't know what to do about it.

none of this is really my concern, because it's all out of my control/realm of my responsibility, but i feel bad for the people that have been there 3+ years. They're probably too committed at this point to back out. I'm just there to learn and collect a paycheck for as long as I can, and I'm fine with that, like I said it's a really comfortable, friendly environment where everyone is pretty much pals.

everyone's been there at least 2-3 years too so that tells me it's not a ****ty place to work for. They overpay I think, so maybe that's why. Idk. It's definitely weird.



I'm actually not too scared about the project mgmt stuff, people seem relieved to have someone just driving the process. I'm scared they'll give me stuff to code and I have no idea how to do it. Luckily they don't have a lot of confidence in my programming ability.

How much of the performance is due to not being ACID compliant?

jmakin,

They have raised a tiny amount of money.

There is no way they are savvy business people in the slightest.

They can easily keep raising enough money to keep the company going indefinitely.

Keep it as a learning experience, and the best learning comes from failure that you don't have to pay for.