Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
ohnoesjmr
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
11 ms
·
61.
▲
by
ohnoesjmr
7y ago
Running hadoop nodes, name nodes, yarn masters. These are massive complicated processes, that cumulatively are usually more complicated and expensive than the spark jobs themselves. At least from the cases I've seen.
62.
▲
by
ohnoesjmr
7y ago
I congratulate the effort, as I always thought that Spark is great, but the fact it was written in Java hinders it quite badly (GC, tons of memory required for the runtime, jar hell (want to use proto3 in your spark job? Good luck)). I do h
63.
▲
by
ohnoesjmr
7y ago
It's total cost of ownership. Spark is a beast, half the resources is taken up by the platform itself.
64.
▲
by
ohnoesjmr
7y ago
Hey, but it's great if you want to piggy-back a botnet on it, with no way to kill c&c: https://www.anomali.com/blog/the-interplanetary-storm-new-ma...
65.
▲
by
ohnoesjmr
7y ago
Presumably this needs it's own IP sockets, which means this needs root to run?
66.
▲
by
ohnoesjmr
8y ago
Can't wait for different language models to be available to download for recognition, so one could have a genuine offline dictation between languages.
67.
▲
by
ohnoesjmr
8y ago
Is it the same proxy that is used in redis enterprise today?
68.
▲
by
ohnoesjmr
8y ago
Why wouldn't it send SNI for the real domain? Also, what's preventing me from claiming some domain I don't actually own, if you are allowed to wedge in any invalid tld and nothing is checked?
69.
▲
by
ohnoesjmr
8y ago
Not sure I understand how this attack works. You ask for cert for foo.com, LE tells you to provide a self-signed cert subjectAlt=foo.bar.acme.invalid. LE resolves foo.bar to some heroku endpoint as the domain lives on heroku, LE connects vi
70.
▲
by
ohnoesjmr
8y ago
Yeah, first few hundred bytes are all zeroes.
71.
▲
Ask HN: Recovering an audio disk ripped in an unknown format
3 points
by
ohnoesjmr
8y ago
|
7 comments
72.
▲
by
ohnoesjmr
8y ago
For the funds that have relations with policits, I think it's mostly getting rid of regulations imposed by the EU.
73.
▲
by
ohnoesjmr
8y ago
I really struggle with this 5 years in... I open my mouth when I sleep, probably because of adenoids (had them removed twice, keep growing back), so I can't use nose pillows/masks, and the full face mask leaks badly as my bottom
74.
▲
by
ohnoesjmr
8y ago
For me, I'd crash straight after work one day a week, I just couldn't last till the evening (I work an office job, so no physical tiredness). Also, I recall driving on a highway was terrible, I'd start nodding off after 10min
75.
▲
by
ohnoesjmr
8y ago
You can't really do that with distinct, as if you have 1 billion distint entries, you essentially have to store all of them to dedup.
76.
▲
by
ohnoesjmr
8y ago
Last time I looked at this, the code in general and code you needed to use it was overly complicated for what it does.
77.
▲
by
ohnoesjmr
8y ago
This is misleading. The title should be "Using s3 as storage and some side thingy that runs every minute as a scheduler".
78.
▲
by
ohnoesjmr
8y ago
I am pretty sure I am not, as I received the message that onboarding is disabled. If you have user requests in the logs, you should queue them up.
79.
▲
by
ohnoesjmr
8y ago
Given you disabled onboarding, is my onboarding request queued somewhere or will I have to comeback and re-request once you got over the hump?
80.
▲
by
ohnoesjmr
9y ago
SNI is still in plain text :( They could just hash the SNI and do matching based on hashes.
81.
▲
by
ohnoesjmr
9y ago
I wonder how this compares with Gogoproto. Nobody in their sane mind uses the standard protobuf package anyway.
82.
▲
by
ohnoesjmr
9y ago
Yeah, I fail to understand the purpose of the dispatcher thingy aswell. This also can run out of memory as he keeps spawning routines indefinately with each request in the dispatcher.
83.
▲
by
ohnoesjmr
9y ago
It's quite poor. Two attempts as a non-english speaker in my normal voice, and 70% british first, then 70% american.
84.
▲
by
ohnoesjmr
9y ago
I wonder if one could abstract tcp on top of udp using this.
85.
▲
by
ohnoesjmr
9y ago
The problem with contexts is that can't just innocently enter one in some if branch where the rest of the code would be scoped. I ended up an OnExit() context that does: if conn is not None: ctx.call(conn.close)
86.
▲
by
ohnoesjmr
9y ago
Had the same, was diagnosed with sleep apnea.
87.
▲
by
ohnoesjmr
9y ago
[spoiler] It unlocks at 10, 50 and 100 factories or something like that, same with drones, 5k, 50k etc. [/spoiler]
88.
▲
by
ohnoesjmr
9y ago
Sounds very cool, but first thing that hit my head: Sounds like a perfect tool to DDoS someone?
89.
▲
by
ohnoesjmr
9y ago
Why not simply full indexes?
90.
▲
by
ohnoesjmr
10y ago
Yeah, the only problem is you can never "un-give" a key, which means you either have to rebuild your cluster under a different key (having to re-give key to everyone else apart from that one guy you want to kick out), or deal with
More ›