orf

55 ms·

931.
by orf 4y ago
Why not randomise VPCs between instances? It’s fairly easy to do with a fixed pool and random selection.
932.
by orf 4y ago
Something must be wrong, at my previous work we were able to achieve a much higher aggregate throughput - are you spreading them across AZ, and are you using a VPC endpoint? We did use multiple VPCs, that might make a difference
933.
by orf 4y ago
Ahh yes, sorry, you are running it on a single instance. That’s capped, but the aggregate throughput across instances can be a lot higher. There are request limits but these are per partition. There is also an undocumented hard cap on list
934.
by orf 4y ago
You won’t be throttled if you partition your data “correctly” and have a reasonable amount of data in your bucket. That’s difficult to do, but more than possible.
935.
by orf 4y ago
The issue as I understand it is that there is no way to encode this information in the Kubernetes token. The pod IAM roles stuff leverages Kubernetes stuff, and the token that’s mounted into the container is a YAML representation of a Kuber
936.
by orf 4y ago
Interesting that they use a single IAM role per client. This will also run into limits (which can be upped), but wouldn’t a better design be to use a single role with a specific STS session name and session tag enforcement policy? You can t
937.
by orf 4y ago
It’s perfectly ok to post that, but don’t expect to not be downvoted for it
938.
by orf 4y ago
I wasn’t aware that this was the situation in the US. > and one that claims untold billions of dollars at US gas stations, and many multiples of that amount at european ATMs I’m interested in why you think this - from everything I can fi
939.
by orf 4y ago
> Not more crazy than giving it at the restaurant or store But why would you let your card leave your sight? I don’t understand. The waiter brings over a payment terminal or you go over to it, then you tap or insert your card. When coul
940.
by orf 4y ago
Why would you give credit card information over the phone to anyone? I’m not in the US but that sounds crazy.
941.
by orf 4y ago
There are similarities in the implementation as well, specifically around content addressable storage and layers. Nix needs reproducible builds as well, which is a limiting factor. What’s nice about Docker is that it’s flexible enough to ac
942.
by orf 4y ago
There’s purity of design/implementation and there’s usage. Often these are inversely related. I like nix - it’s probably the right direction, but just compare the UX to Docker. Comparatively Docker absolutely nailed it.
943.
by orf 4y ago
I think you are letting your specific feelings and head-canon about purity be mistaken for solid technical arguments. If you’re sshing to boxes, editing things by hand and slinging ad-hoc commands around then your frame of reference is so f
944.
by orf 4y ago
At one point we had a ~560tb EFS disk that ran a variety of mixed workloads (large and small files). It was untenable - raw reading/writing IO is OK, but metadata IO hits a brick wall and destroys the performance of the whole disk for
945.
by orf 4y ago
I also maintain a huge shell history, but in my experience fzf-ing through it becomes a lot harder when it’s terse and unreadable. But sure, there’s a tension between readability and brain-to-command throughput that’s always been there. The
946.
by orf 4y ago
Cool but that’s as opaque and hard to understand as sacrificing a goat to order ls output. It’s of course possible to do with existing shells, but it’s impossible to argue that it’s a clear
947.
by orf 4y ago
Thanks! I was aware of the Aphyr series, but he said Richard also had a series.
948.
by orf 4y ago
SSDs don’t wear from reading, only from writing. Assuming a sensible, somewhat linear layout using mmap to map the weights would give you the ability to load a lot in memory, with potentially a fairly minimal page-in overhead
949.
by orf 4y ago
This might be naïve, but couldn’t you just mmap the weights on an apple silicon MacBook? Why do you need to load the entire set of weights into memory at once?
950.
by orf 4y ago
You can do this quite nicely with serde in Rust. It’s elegant. There’s still allocations when needing to unescape values and expand the scratch space, but otherwise it’s all zero copy. You can parse at over a gigabyte a second from stdin wi
951.
by orf 4y ago
Death by a thousand cuts. Let’s take JSON parsing as an example. In the simple case a value like “foo” requires no copying during parsing - it can be parsed with zero copies. But escaped values like “foo\”bar\”” cause issues. The actual str
952.
by orf 4y ago
Can you link them? I couldn’t find them on his blog.
953.
by orf 4y ago
Yes, I use this a lot and it’s fantastic. Works pretty flawlessly, is fast and super simple to set up.
954.
by orf 4y ago
A simple point class looks like this: class Point { has [qw/x y/] => ( default => sub { 0 } ); sub clear ($self) { $self->@{ 'x', 'y' } = ( 0, 0 ); } } What?
955.
Show HN: structured-ripgrep – Ripgrep over structured data(github.com)
15 points by orf 4y ago|2 comments
956.
by orf 4y ago
I guess it’s kind of sweet and innocent to have such faith that a group of disparate people would act against their own self interest for only a nebulous future “common good” outcome, and that would be sparked by a comment on an internet fo
957.
by orf 4y ago
I haven’t really seen many R2 performance benchmarks, but 1s+ for retrieval is nuts. Is that time to first byte or total time? On what average object size?
958.
by orf 4y ago
Is it ironic that his comments are harder to read (on mobiles) precisely because of his weird self-imposed word wrapping? I get it. His arguments make sense because they don’t want to bake in a specific “start block quote character”, and wi
959.
by orf 4y ago
I’m interested in seeing you reply to the other commentators about the cops who attacked Floyd. How do you defend this position?
960.
by orf 4y ago
The answer to that question depends really on the context. At the end of the day the borrow checker is there to help you. If a function takes ownership of a value, like a string, it generally means that it wants exclusive access to the memo

More ›