Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
puzzle
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
23 ms
·
241.
▲
by
puzzle
8y ago
Now the real question becomes: how often were they getting phished before the new policies? Knowing Google, there's no way they will answer THAT before another decade.
242.
▲
by
puzzle
8y ago
I thought that HDMI and DisplayPort used compression, too, sometimes. Even then, your point about generic computing still stands.
243.
▲
by
puzzle
8y ago
Reminds me of the story of the heroic woman in Arkansas who, during the same period, helped hundreds of victims living their last days on Earth: https://m.arktimes.com/arkansas/ruth-coker-burks-the-cemeter...
244.
▲
by
puzzle
8y ago
Maybe you read the opposite, i.e. that GCP's KVM virtual machines run inside Borg containers?
245.
▲
by
puzzle
8y ago
Yeah, but with Borg you can have many of them, all directly accessible with one click from the Borg job status page: e.g. a general status, a view of the work queue if it's something like a crawler, incoming and outgoing RPCs, memory&#
246.
▲
by
puzzle
8y ago
They're sometimes interactive HTTP endpoints, not metrics, although metrics are served by one of them (/varz, described in the SRE book). They're informally called z pages and are the inspiration behind the /debug handle
247.
▲
by
puzzle
8y ago
> Google is still relying on Borg for its internal production infrastructure, but I can tell you it’s not because of the design of Borg interfaces! No matter how spartan, the Borg status pages are more helpful than most Kubernetes UIs ou
248.
▲
by
puzzle
8y ago
They mentioned that Kubernetes runs some workloads and they are probably using Docker for that, like GKE does (17.03, I think). No way they would use it for Borg.
249.
▲
by
puzzle
8y ago
LOAS is how Stubby (gRPC) connections on the internal network are secured, looks a bit like mTLS and is partially being open sourced as ALTS: https://cloud.google.com/security/encryption-in-transit/appl... GFE (Go
250.
▲
by
puzzle
8y ago
There's more: Cos and the open source Vanadium, John Hanke and Ingress/FieldTrip, Niniane and Lively, Hartmut and the quantum lab, Craig N-M and google.org, etc. But you left out the largest retention side project: Jeff Dean and B
251.
▲
by
puzzle
8y ago
Another reason postmortems can be vague is that, for each detail you add, you might need to add even more background details. Then it becomes a recursive problem. In this case, the "configuration change" could be a feature flag in
252.
▲
by
puzzle
8y ago
Sometimes bugs are triggered by very specific data or queries (aka "queries of death"). GFEs have many tests, so I'd go for the subtle bug.
253.
▲
by
puzzle
8y ago
I wasn't around at the time, but I suspect it's mainly because the first datacenters were on the West Coast. See e.g. the story at http://www.dodgycoder.net/2013/02/googles-fiber-leeching-cap...
254.
▲
by
puzzle
8y ago
Ten years ago it was called jokingly GST: not Gulf Standard Time, but Google Standard Time. I don't remember outages, but the monitoring graphs when switching to DST, where plots go back in time, did look very funky. Something like thi
255.
▲
by
puzzle
8y ago
Yeah, Google hired Robert Geisberger...
256.
▲
by
puzzle
8y ago
At some point, there was also this undocumented feature where it fetches not just the directions, but a little extra data so it can reroute you on the device if you miss a turn and your phone has gone offline. It only works for small detour
257.
▲
by
puzzle
8y ago
Yes, only Russian agents are indicted this time, but someone who might be Roger Stone makes a brief appearance (#44) in the document, too, probably to be targeted in future legal action.
258.
▲
by
puzzle
8y ago
No, unless you have a special setup where one process on the whole machine can use AVX. That might make sense for special, controlled environments. What I meant is that there are more registers to shuffle around at context switch time, when
259.
▲
by
puzzle
8y ago
You also have penalties from context switches, although you can reduce their impact by performing them in a lazy fashion. AVX512 is even worse, of course.
260.
▲
by
puzzle
8y ago
Google has more POPs, search clusters everywhere, etc. I was going to say that they also use QUIC, but the poster mentions using Firefox, which I don't think supports it yet.
261.
▲
by
puzzle
8y ago
I loved this language. I still have all the papers somewhere. Unfortunately, the announcement of Java killed the few remaining chances it had of outliving General Magic.
262.
▲
by
puzzle
8y ago
Yeah, I know about OS/400 using the extra bytes on hard drives for the pointer tags (not a checksum), but what I had in mind were data CDs, whose canonical sector size was 2048.
263.
▲
by
puzzle
8y ago
Yeah, you're right, they wanted 512 instead of 2048 bytes per sector to use the CD as a HD. But I think that at least one of the Unix vendors shipped the install media as 520 byte sectors, with 512 bytes of data and 8 bytes of checksum
264.
▲
by
puzzle
8y ago
Not too bad. I was expecting the first shaving step to be about the OS media requiring 520 byte sectors and thus needing a "fancier" CD writer/reader. Maybe that was Sun, not SGI?
265.
▲
by
puzzle
8y ago
So true. Inside Google, it used to be the case that every N months someone had to figure how to build their Python daemon with Blaze (Bazel) using tcmalloc. It was painful and I hope it's easier nowadays. Without it, after a few weeks
266.
▲
by
puzzle
8y ago
I was at Google and they did buy offsets from 2007 until after I left. Not for Oregon, sure, but definitely for places like the Carolinas (our good old friends at Duke Energy...). It was only in 2017 that they covered 100% of corporate/
267.
▲
by
puzzle
8y ago
Like the other poster, I was at Google during those days, but I do not want to say too much. I never dealt with Apple myself, but I knew people who did, for both Maps and YouTube. None of them were particularly delighted. The All Things D a
268.
▲
by
puzzle
8y ago
ISTR it's the latter, so that users don't need to enter username and password (and maybe 2FA) five times because they install five different Google apps.
269.
▲
by
puzzle
8y ago
Originally, they couldn't license the actual data, either, because at least the road network wasn't Google's to license. It came from TeleAtlas (after Nokia had bought NavTeq), Zenrin for Japan and a few more. Things were tri
270.
▲
by
puzzle
8y ago
Apple never got data in the original deal, either. They just proxied tiles and search requests. That story is cool, but it's missing at least an additional plot twist that "people familiar with Google's thinking" cautiou
More ›