malisper

23 ms·

271.
by malisper 7y ago
There's a trade off here in that replicating data decreases the read load but increases the write load. If you have a chunk hot with writes, increasing the replication will make things worse, not better.
272.
by malisper 7y ago
I can speak a bit to this. I previously lead the database team at Heap where we had one of the largest deployments of Citus with ~1PB of data. Citus and Aurora Serverless are pretty different. The only thing in common is that they are both
273.
by malisper 7y ago
I'm not aware of any general work. I initially tried to directly port Paul Graham's interpreter into SQL. I immediately ran into problems because it wasn't possible to map the recursive function calls into SQL. After thinking
274.
by malisper 7y ago
Speaker here. If you liked the talk, you may also like the other talk I've given on generating fractals with SQL[0]. You can also find it blog post form[1]. Originally, before I had given the fractal talk, I had wanted to make a compil
275.
Writing an Interpreter in SQL for Fun and No Profit [video](youtube.com)
55 points by malisper 7y ago|11 comments
276.
by malisper 7y ago
Hi. Creator here. I developed Perfalytics out of my frustration with analyzing Postgres performance. I've been optimizing Postgres full time for the last three years and I still find it difficult to get a good understanding of why cert
277.
Show HN: Perfalytics – Postgres Performance Made Easy(perfalytics.com)
2 points by malisper 7y ago|1 comments
278.
by malisper 7y ago
I like what it's trying to do, but I disagree with the approach. I have a few issues with the implementation: #1 Dexter relies solely on the Postgres cost model. Dexter works by using HypoPG[0] to create a hypothetical version of every
279.
by malisper 7y ago
I have a number of thoughts on this. A common claim I've heard is that query optimizers are better at producing efficient query plans than all but the best programmers. I can't find any concrete studies on this, but it's plau
280.
by malisper 7y ago
I can definitely relate to the struggle of debugging slow SQL queries. I've been optimizing Postgres for the last three years for my day job. There's nothing like spending hours trying to determine why a query is slow, only to dis
281.
by malisper 7y ago
> My current understanding is that json is faster and bigger than jsonb. FYI, it's the other way around. When you use the json type, Postgres validates it and stores the raw string. Operations on json are very slow. Whenever you ext
282.
Generating Fractals with Postgres: Self-Similar Fractals(malisper.me)
1 points by malisper 7y ago|0 comments
283.
by malisper 7y ago
I was trying to think of an idea to submit to !!con West: http://bangbangcon.com/west/ My original idea was to write a compiler from a programming language to SQL. I was trying to think of what would be cool examples o
284.
by malisper 7y ago
Author here. I had a great time figuring out how exactly to write the queries in the blog post. If you have any questions, feel free to ask.
285.
Generating Fractals with Postgres: Escape-Time Fractals(malisper.me)
69 points by malisper 7y ago|8 comments
286.
Generating Fractals with SQL Queries(malisper.me)
1 points by malisper 8y ago|0 comments
287.
by malisper 8y ago
Agreed. The business models and risks between these companies are completely different. The main commonality I was highlighting was that these business have high upfront costs and make the money back over a long period of time. Because of t
288.
by malisper 8y ago
I'm seeing a lot of people in this thread balking at the $2B in losses. The thing is that aggregate profit is a terrible metric for measuring companies that have a recurring revenue model with high upfront costs (as WeWork, Uber, and m
289.
by malisper 8y ago
The update is at the bottom: > Update: After spending about a year on the BYOT experiment, we’ve decided to sunset it. In short, it didn’t work; we didn’t hire any teams as a result. > There were a few reasons for this, which we learn
290.
Bring Your Own Team Update: BYOT Is Being Sunset (2017)(stripe.com)
1 points by malisper 8y ago|1 comments
291.
by malisper 8y ago
To be fair, an i3.4xl has 10x the disk and 4x the memory as a c5d.4xl at 1.5x the cost. If you are optimizing for GB/cost, i3s are still a reasonable choice. There's also the i3.metal which has no hypervisor. I'm curious how
292.
by malisper 8y ago
We store two copies of every piece of data on two different machines. When a single machine goes down, we have code for spinning up a new machine and restoring the data that was on the machine. Over the course of a month, we usually have ab
293.
by malisper 8y ago
We are running vanilla ZFS-on-Linux. We don't use snapshots for backups as the Postgres backups are more convenient. Postgres provides point in time recovery, which is useful. There are also tools like wal-e for automatically writing a
294.
by malisper 8y ago
When we first launched the i3 instances, i3.metals weren't available. We've been wanting to do experiments with i3.metals, but we've been unable to get confirmation that our reservations will transfer over. Until we know that
295.
by malisper 8y ago
Hi. I'm one of the database engineers at Heap. This is a good question. There are several reasons why we use EC2. First of all, I will say I love RDS as a product. We actually do use RDS for a number of our services. We use Postgres on
296.
by malisper 8y ago
I agree. The author of the post didn't require that the general's attack at the same time. They only had to reach consensus on whether to attack or retreat. That's what spooneybarger meant by: > The article incorrectly def
297.
by malisper 8y ago
Normally in the two generals problem you assume that if you send a message repeatedly for an indefinite amount of time it will eventually be delivered. This assumption makes the problem the author described possible, but the actual two gene
298.
by malisper 8y ago
Deciding between a time and an action does matter. The big difference is when deciding on an action, each general only needs to eventually learn the action to take. When learning a time, each general needs to know the time to attack and t
299.
by malisper 8y ago
Hmm... I don't think the comparison says anything about Hadoop vs Bash. The programs they are running are completely different. In that article, the author runs a regular expression that looks lines that start with "Result" i
300.
by malisper 8y ago
ActorDB sounds like an interesting piece of software. It looks like it's good for simple applications that have large amounts of data. Some thoughts I have based on the docs: ActorDB has no concurrency at the actor level. This makes it

More ›