Hacker News: leg

New comment by leg in "Colossus for Rapid Storage"

leg — Thu, 10 Apr 2025 21:45:09 +0000

Anywhere Cache shines in front of a multi-regional bucket. Once the data is cached, there's no egress charges and there's much better latency. This is great for someone who looks for spot compute capacity to run computations anywhere in the multi-region. It will also improve performance in front of regional buckets but as a cache, you'll see the difference between hits and misses.

Rapid Storage will have all of your data local and fast, including writes. It also adds the ability to have fast durable appends, which is something you can't get from the standard buckets.

New comment by leg in "Colossus for Rapid Storage"

leg — Thu, 10 Apr 2025 17:58:17 +0000

Anywhere Cache and Rapid Storage share some infrastructure inside of GCS and both are good solutions for improving GCS performance, but Anywhere Cache is an SSD cache in front of the normal buckets while Rapid Storage is a new type of bucket.

(I work on Google storage)

New comment by leg in "Colossus for Rapid Storage"

leg — Thu, 10 Apr 2025 17:25:21 +0000

There's a detailed blog post about Rapid Storage now available, see https://news.ycombinator.com/item?id=43645309

(I work on Google storage)

New comment by leg in "Railroad Tycoon II"

leg — Mon, 13 Jan 2025 13:08:29 +0000

I've played 100s of hours of if under Proton on Linux. Works great.

New comment by leg in "About synchronous disk replication"

leg — Sat, 31 Aug 2024 19:00:55 +0000

Many enterprise storage systems have the durability/availability tradeoff like these replicated disks when replicating outside of a single datacenter. (Oracle calls it "max availability": try to synchronously replicate, but if the remote side is offline, allow transactions to commit.) Real world banks run on these sorts of systems.

Users don't continuously check replication status. They rely on it being synchronous almost all the time.

3 way quorum replication is great, but you then need to send to more data centers, potentially affecting performance. There's a tradeoff.

(I work on GCP storage)

New comment by leg in "About synchronous disk replication"

leg — Sat, 31 Aug 2024 18:56:20 +0000

Regional Persistent Disk was in beta in 2019. Usability hiccups and other annoyances meant it only GA'd in 2023, but it's been used under CloudSQL for quite a while.

(I work on GCP storage)

Synthesized Google storage I/O traces now available to the community

leg — Tue, 25 Jun 2024 17:41:30 +0000

Article URL: https://cloud.google.com/blog/topics/systems/synthesized-google-storage-io-traces-now-available-as-open-source/

Comments URL: https://news.ycombinator.com/item?id=40791290

Points: 2

# Comments: 0

New comment by leg in "C++ Exit-Time Destructors"

leg — Sun, 17 Mar 2024 21:10:05 +0000

std::quick_exit() also works, though you can decide if it's worthwhile to allow parts of the program to register functions to be called at quick exit time.

New comment by leg in "Citigroup Plans to Cut 20k Jobs"

leg — Fri, 12 Jan 2024 16:54:59 +0000

Linked investment account counts. You just need $2500 in a checking account, the rest can be in whatever ETF or stock you'd like.

Granular instance sizing for Cloud Spanner

leg — Tue, 31 May 2022 21:43:16 +0000

Article URL: https://cloud.google.com/blog/products/databases/use-spanner-at-low-cost-with-granular-instance-sizing

Comments URL: https://news.ycombinator.com/item?id=31575320

Points: 2

# Comments: 0

New comment by leg in "Google's infamous internal 2010 “I just want to serve 5TB” video now public"

leg — Tue, 02 Nov 2021 17:14:30 +0000

It is true that Borgmon readability went away due to this video. It wasn't Sergey, it was an eng director.

High Durability Options for Compute Engine Workloads

leg — Fri, 20 Aug 2021 02:12:17 +0000

Article URL: https://cloud.google.com/blog/products/compute/high-durability-persistent-disk

Comments URL: https://news.ycombinator.com/item?id=28241647

Points: 1

# Comments: 0

Colossus a peek into Google’s scalable storage system

leg — Mon, 26 Apr 2021 15:08:54 +0000

Article URL: https://cloud.google.com/blog/products/storage-data-transfer/a-peek-behind-colossus-googles-file-system

Comments URL: https://news.ycombinator.com/item?id=26943310

Points: 1

# Comments: 0

Creating a persistent disk clone from a source disk

leg — Fri, 10 Jul 2020 18:35:45 +0000

Article URL: https://cloud.google.com/compute/docs/disks/add-persistent-disk#source-disk

Comments URL: https://news.ycombinator.com/item?id=23794757

Points: 1

# Comments: 0

Using a petabyte-scale database's monitoring tools to make art

leg — Tue, 05 May 2020 19:00:48 +0000

Article URL: https://medium.com/google-cloud/using-a-petabyte-scale-databases-monitoring-tools-to-make-art-8aa8eb17234

Comments URL: https://news.ycombinator.com/item?id=23083524

Points: 2

# Comments: 0

Dynamic Hybrid-SMR: an OCP proposal to improve data center disk drives

leg — Mon, 13 Nov 2017 15:45:03 +0000

Article URL: https://blog.google/topics/google-cloud/dynamic-hybrid-smr-ocp-proposal-improve-data-center-disk-drives/

Comments URL: https://news.ycombinator.com/item?id=15687235

Points: 3

# Comments: 0

New comment by leg in "Decommission of Cyrus Email"

leg — Sun, 07 May 2017 13:52:55 +0000

Yeah, John is great. His Unicode handling in Cyrus was way ahead of it's time.

I left CMU & the project 14 years ago. It's nice to know that some code lives on.

New comment by leg in "In search of a simple consensus algorithm"

leg — Fri, 17 Mar 2017 15:57:12 +0000

etcd didn't implement leases; it just assumed that the last Raft election was still good.

You can elect a leader for a set period of time (say, 10 seconds) and serve strong reads for a lesser period of time (5 seconds) if you have reasonable assumptions of how good your local oscillators work and avoid jumps. If you don't trust your local clock to any level of accuracy, why do you trust your local CPU?

Cockroach DB does this correctly. https://github.com/cockroachdb/cockroach/blob/master/docs/de...

New comment by leg in "In search of a simple consensus algorithm"

leg — Fri, 17 Mar 2017 15:06:30 +0000

The performance penalty is on reads. If you elect a leader with a timeout you can do strong reads without re-establishing consensus.

New comment by leg in "Inside the Obama Tech Surge as It Hacks the Pentagon and VA"

leg — Wed, 20 Jul 2016 21:09:14 +0000

Except that state and local governments are mostly more corrupt, more poorly run, and less well supervised by the voters.