puzzle

34 ms·

271.
by puzzle 8y ago
That is not new. And it's still not enough, because it's not impossible to reconstruct someone's route if they drive in areas where they are the only ones (or among the few) to send segment data. That's why you need to d
272.
Kubernetes 1.11(github.com)
98 points by puzzle 8y ago|5 comments
273.
by puzzle 8y ago
You might be joking, perhaps not too much, but at some point Google used NFS (see https://wiki.ubuntu.com/AndrewPollock2 ), then replaced it with something else: https://github.com/google/clusterfuzz-too
274.
by puzzle 8y ago
More questions: Does it support extended attributes? You'd imagine so, but you'd have also expected NFSv4... Will it ever support SMB? Can you set per-UID quota limits, both in terms of bytes and IOPS/disk time?
275.
by puzzle 8y ago
Well, the poster probably meant whatever GFS incarnation is around these days (Colossus is v2 or v3, depending on how you look at things). In the end, almost everything at Google is backed by "GFS": Bigtable, Blobstore, GCS, Spann
276.
by puzzle 8y ago
I'd normally recommend books like Google's SRE one, but at least in this case it glosses over the detail of where GFEs tend to live: https://landing.google.com/sre/book/chapters/production-envi... I
277.
by puzzle 8y ago
Deepmind has at least one example of saving Google millions: https://deepmind.com/blog/deepmind-ai-reduces-google-data-ce...
278.
by puzzle 8y ago
There's a difference between a data center and a POP. The former is very large. The latter is usually a rack or more in someone else's facilities. When you try to reach Google or Facebook, you rarely hit their data centers directl
279.
by puzzle 8y ago
It also deters blackmail from other parties, if the relationship is not public.
280.
by puzzle 8y ago
The logging service numbers are easily explained: the remote server might have transient failures, so the forwarder will cache stuff in memory (the alternative is to just stop reading the logs, but then you risk losing messages if the pod d
281.
by puzzle 8y ago
Were the system pods using all that memory or just reserving it? It's not straightforward to scale them, because the node might run just your tiny Rust server or 20 high traffic web apps. You don't want the log agent to keel over
282.
by puzzle 8y ago
I sent feedback from the YouTube app while on a blocked video. It gives you the options to contact them via chat, phone or email (the last one is composed all within the app, not in your email client).
283.
by puzzle 8y ago
Yeah, it's not like Uber did not hire Waleed Kadous away from Google, where he had led Android and indoor positioning, which make plenty of use of fused location. I bet he was involved in recruiting these folks (he approached someone I
284.
by puzzle 8y ago
I never cracked disk protections, but I did sometimes reverse engineer games to figure cheat codes, etc. One EA game I looked at was the Will Harvey's Zany Golf for the Amiga. I had read somewhere that there was a secret level (the mad
285.
by puzzle 8y ago
The original Codemasters was one of a kind. Rock Star Ate My Hamster was a lot of juvenile fun.
286.
by puzzle 8y ago
As a data point, GV decided twice not to invest in Theranos, due to lack of transparency, a peculiar board mix and a sketchy experience in the field. They actually sent someone to a Walgreens. Later they happily poured lots of millions into
287.
by puzzle 8y ago
Yes, I mention it above. It's an Agenda VR3, a Linux PDA from 2001. I still have one of those. The CPU is from 2000.
288.
by puzzle 8y ago
When I said "for some reason", I was trying not to go into speculation. But yeah, ARM definitely attacked them in the embedded market from the lower end. And Motorola, transitioning from 68k to PPC, from the high end (PPC was very
289.
by puzzle 8y ago
The Linux Agenda VR3 ran on two AAA batteries and used a 66MHz NEC VR4181. That's a 32/64bit MIPSIII chip. The VR3 only used it in 32bit mode, but you could take over the system and enable 64bit mode. It was little-endian only, un
290.
by puzzle 8y ago
Probably nowhere. :-( MIPS was, almost twenty years ago, perhaps the first architecture that had 64-bit capable chips you could power with two AAA batteries (along with RAM, a screen and the rest of the system). Talk about squandered opport
291.
by puzzle 8y ago
HTTP/1.1 to the backend? How about HTTP/2 so people can run gRPC services straight through without translations?
292.
by puzzle 8y ago
The Amiga chipset used DMA very extensively, so the CPU's memory access speed is not as important as on other machines. There were other architectural constraints, like the fact that the CPU had to compete with such chips for access to
293.
by puzzle 8y ago
Beyond the rethoric, the substance is not very different from RMS' arguments against TPMs, which must be some 15 years old now. You'll have a hard time buying a device today without such a chip.
294.
by puzzle 8y ago
It's the new TPM? The old one had a (now patched) firmware bug, which newer Chromebooks like the Pixelbook don't seem to be vulnerable to: https://www.chromium.org/chromium-os/tpm_firmware_update Edit: this i
295.
by puzzle 8y ago
He's completely right that the data is messy and sometimes makes no sense at all. Even a passenger waiting for a train, if paying attention, will notice trains appear, disappear and "jump around" on the countdown clocks.
296.
by puzzle 8y ago
Yeah, it was safe to assume they would announce something related at re:Invent, given that they had joined the CNCF a few months before, after simply ignoring Kubernetes for years.
297.
by puzzle 8y ago
I see. Perhaps latency can be an alternate metric to divine capacity/maxqps, because you could have a backend that is not using a lot of CPU, but is still very slow, because it's just waiting on the local DB (which is the one melt
298.
by puzzle 8y ago
How do you set up capacity for each cluster? You'll reach a point where one cluster will be overloaded; then, it makes more sense to send some requests to a cluster in another region, rather than melting down the local one. GSLB suppor
299.
by puzzle 8y ago
Isn't it funny how Google announces free GKE stuff the day before EKS was announced AND made generally available? :-) Like Jobs issuing a press release that the App Store had hit 3B downloads the morning of the Nexus One launch.
300.
by puzzle 8y ago
A more accurate point is not necessarily that Google will keep some more nodes up, but that it can fit small workloads like a Kubernetes master into existing nodes that run other stuff and have bits of spare capacity. Borg's resource e

More ›