Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
dikei
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
23 ms
·
241.
▲
by
dikei
5y ago
In Windows, people willingly download shady binary from piracy sites all the time to avoid paying for software. I think you underestimate the incentive of free (beer) stuffs.
242.
▲
by
dikei
5y ago
Unless your own benchmarks prove hash indexes to have major benefits for your queries, stick with the default B-tree index: you never know what queries you'll need in the future.
243.
▲
by
dikei
5y ago
Depending on the database, changing the storage layer might require changing almost everything: the query planner, optimizer and executioner intrinsically need to know how storage works to make queries efficient.
244.
▲
by
dikei
5y ago
>> There can only be one main node at one time This is how most distributed system solve the consensus problem: the nodes automatically run leader election algorithm so that eventually only one leader remains. >> (Secretly not s
245.
▲
by
dikei
5y ago
Distributed systems are not limited to distributed computing: once you start breaking your monolith application into multiple services, you have yourself a distributed system.
246.
▲
by
dikei
5y ago
There are no conflict: key-value is an interface, segment files is an implementation. You can implement a key-value using segment files too.
247.
▲
by
dikei
5y ago
Most application never reaches the scale that requires sharding. I'll just use the quote from the notion blog as arguments against automated sharding solution like Vitess. >> Besides introducing needless complexity, an underrated
248.
▲
by
dikei
5y ago
I don't think there will be a monopoly for SpaceX, at least not a global one. It's a matter of nation sovereignty / media control: governments need local service providers that they can controlled. There's no way China w
249.
▲
by
dikei
5y ago
We use Pgbouncer in in our Apache Airflow cluster. During busy time, Apache Airflow can launch hundreds of workers, each of them need at least 1 connection to Postgresql to save its result. Because each worker run in its own process due to
250.
▲
by
dikei
5y ago
Tomtom is selling data to businesses that require raw map data to build their own apps like ride haling service, delivery service, etc..
251.
▲
by
dikei
5y ago
For existing table, HypoPG is a very nice extension that can help you check if your index is helpful before actually creating it. https://github.com/HypoPG/hypopg
252.
▲
by
dikei
5y ago
> In practice, you should typically only send write traffic to 1 galera node Yes, we did that too, it's the only way to reduce the deadlock and certification failure with Galera. However, it won't prevent a single `ALTER TABLE`
253.
▲
by
dikei
5y ago
Only people who haven't run MySQL multi-master HA at scale praise it. In reality, it's a world of pain, where one single seemingly innocent query can take down the entire cluster. Speaking from operation experience with MariaDB Ga
254.
▲
by
dikei
5y ago
I concur. Furthermore, the pricing if often opaque: you know the tools are expensive when they always want you to contact sales for a quote.
255.
▲
by
dikei
5y ago
It's also easy to over-tune your JVM because there's so many options. For modern JVM, I've found that only setting a large enough heap size is good enough for the majority of the software.
256.
▲
by
dikei
5y ago
Not trying to belittle you or anything, but I have a hard time understand why someone would use Go without Goroutine and Channel.
257.
▲
by
dikei
5y ago
> If the end goal is standards compliance, the outcome should be the same. The problem is Google is using Chromium dominant market share to push some questionable "features" into standards. > The fact that Google did such a
258.
▲
by
dikei
5y ago
Firefox used to be extremely extensible: there were some software based on Gecko and Xulrunner. However, they all died out after Chromium and Electron took over.
259.
▲
by
dikei
5y ago
SeaweedFS sounds very interesting indeed, though I'm a bit concern about the complexity. For example, there are so many Filer metadata backend options that it's difficult for me to pick one. I think the document should have a clea
260.
▲
by
dikei
5y ago
No, object store works better with a small number of large object than with a large number of small object. 1kb object would be terribly inefficient if you have too many of them. It's slow even with regular file system, copy 1 million
261.
▲
by
dikei
5y ago
In my opinion, avoid ASN.1 if you can. There's a reason why all the cool companies invented their own serialization formats: Google's Protobuf, Facebook's Thrift, etc.. even when ASN.1 had been an international standard for y
262.
▲
by
dikei
5y ago
Yeah, those are the easy one. I worked in Telecom, and dealing with unaligned PER is a PITA.
263.
▲
by
dikei
5y ago
Can't blame him. ASN.1 is one of the most complicated binary formats, with so many encoding rules that there's no free decoder that can process them all.
264.
▲
by
dikei
5y ago
>> I don't know what it is, but maybe the Chinese army deploys so infrequently (almost never?) that they place zero priority on their MREs. I don't know about the Chinese, but my poor army doesn't have an official MRE.
265.
▲
by
dikei
5y ago
I remember Steve video about China's MRE where he had stomach problem after eating. Eat some biscuit from WW1-era: totally fine, eat some chow mein from PRC's MRE: instant diarrhea
266.
▲
by
dikei
5y ago
Duh, if you need a wallet, go get a new one. I'm not giving you mine /s
267.
▲
by
dikei
5y ago
Well, you can rotate any wallet, if you're willing to pay: just create a new wallet, pay the transaction fee and transfer your money over to the new address. It's kinda like with full disk encryption, you can change the key-encryp
268.
▲
by
dikei
5y ago
Also in the pages that he links, they explicitly warn against password rotation. >> https://github.com/OWASP/ASVS/blob/master/4.0/en/0x11-V2-Aut... >> Note: Passwords are not to hav
269.
▲
by
dikei
5y ago
In most cases, transaction logs/write-ahead logs will contain the return value of non-determistic functions like created_at or random uuid, instead of the function call.
270.
▲
by
dikei
5y ago
I would expected them to have point-in-time recovery, especially from the last time when they accidentally dropped the production database.
More ›