2026-04-16T16:44:42Z https://yabu.me Marc Brooker https://yabu.me/npub1tmlvx33fj5fykqfvupx76fxprlrew9jjgu7cvk04puxscgzgfv8ss5tztd https://fediscience.org/system/accounts/avatars/109/299/638/594/444/863/original/e2a3b0fced45b713.jpg https://fediscience.org/system/accounts/avatars/109/299/638/594/444/863/original/e2a3b0fced45b713.jpg https://yabu.me/nevent1qqsppgms3pcrp7rkf9kx8mkkg6ak6fhe9jacewpy05pzvnaa6krr3rszyp00as6x9x23yjcp9nsymmfycy0u09ck2frnmpje758s6rpqfp9s75cwvwm In reply to <a href='/nevent1qqspwf35fyhszz27y3cl4h4fshf432gnzth0373a5wue7lv8kqra78smqa2mt'>nevent1q…a2mt</a><br/>_________________________<br/><br/>As far as I know, this issue has existed in PostgreSQL since the introduction of MVCC more than two decades ago. It affects most clustered installations of PostgreSQL. We are working with the PostgreSQL community on a fix.<br/><br/>This issue doesn&#39;t affect Amazon Aurora Limitless Database, or Aurora DSQL. Both of these databases use different methods for ordering transactions, which provide a consistent ordering across replicas. 2025-05-03T16:50:16Z https://yabu.me/nevent1qqspwf35fyhszz27y3cl4h4fshf432gnzth0373a5wue7lv8kqra78szyp00as6x9x23yjcp9nsymmfycy0u09ck2frnmpje758s6rpqfp9s7pkxwft Recently, one of our testing tools at AWS found an issue with transaction visibility in PostgreSQL, where the order that transactions become visible differs between the primary and replicas. Sergey Melnik dives deep on the AWS Database Blog: <a href="https://aws.amazon.com/blogs/database/understanding-transaction-visibility-in-postgresql-clusters-with-read-replicas/">https://aws.amazon.com/blogs/database/understanding-transaction-visibility-in-postgresql-clusters-with-read-replicas/</a> 2025-05-03T16:49:54Z https://yabu.me/nevent1qqsvzzzn5yetd6k3j8g3ar4f0365lkja8dsr8f4dzpl7gu3secsln5qzyp00as6x9x23yjcp9nsymmfycy0u09ck2frnmpje758s6rpqfp9s7lj8y3x In reply to <a href='/nevent1qqsrksfdy52llkl4c0kw0pums8z7kjwfcku4q6jszwcjs4zen89w5yqjl6ce9'>nevent1q…6ce9</a><br/>_________________________<br/><br/>TLA+ as a correctness oracle for deterministic (or even non-deterministic) simulation testing makes a ton of sense to me.<br/><br/>As you say, a big part of that is driving folks to think about what correctness means to them.<br/><br/>In DSQL testing, for example, we&#39;ve used &#34;do what Postgres does&#34; as a correctness oracle for a lot of the non-distributed surface area, just because that&#39;s what many of our customers care about. 2025-04-28T17:04:17Z https://yabu.me/nevent1qqsr9jkumcck5guc7z0mh253899kvawcm0w0x22wwt28kt3wh6yek9czyp00as6x9x23yjcp9nsymmfycy0u09ck2frnmpje758s6rpqfp9s7645utk In reply to <a href='/nevent1qqsfe6ymdtj49suvpxjtp5w3r38ge3jvnhxqvkulxvdggtf6pv7he5gca8656'>nevent1q…8656</a><br/>_________________________<br/><br/>100s for a single primary key, although we&#39;ll improve that over time. 2024-12-11T17:03:03Z https://yabu.me/nevent1qqsgyjfrlm73ngl5dea7kxw2ymtlt74v478adfllc7h6k7chlhemp4gzyp00as6x9x23yjcp9nsymmfycy0u09ck2frnmpje758s6rpqfp9s7edztst My second blog post on Aurora DSQL, with a deep dive on how reads work: <a href="https://brooker.co.za/blog/2024/12/04/inside-dsql.html">https://brooker.co.za/blog/2024/12/04/inside-dsql.html</a> 2024-12-04T17:11:17Z https://yabu.me/nevent1qqsyfrhfu8jsrtt6zhqtwxjzvxduq23paa2dthdv7d3chl7yjncm96gzyp00as6x9x23yjcp9nsymmfycy0u09ck2frnmpje758s6rpqfp9s7fesu7u Rust Foundation Collaborates With AWS Initiative to Verify Rust Standard Libraries <a href="https://foundation.rust-lang.org/news/rust-foundation-collaborates-with-aws-initiative-to-verify-rust-standard-libraries/">https://foundation.rust-lang.org/news/rust-foundation-collaborates-with-aws-initiative-to-verify-rust-standard-libraries/</a><br/><br/>Verify the Safety of the Rust Standard Library <a href="https://aws.amazon.com/blogs/opensource/verify-the-safety-of-the-rust-standard-library/">https://aws.amazon.com/blogs/opensource/verify-the-safety-of-the-rust-standard-library/</a> 2024-11-21T00:06:11Z https://yabu.me/nevent1qqsxe86qcnpyt75zxkuuwvl8tvnzusyg0r3etz86a2j3fcgng6mfpqgzyp00as6x9x23yjcp9nsymmfycy0u09ck2frnmpje758s6rpqfp9s7rdsf28 Great new piece from Byron Cook about automated reasoning at AWS, and how we&#39;re finding it not only allows us to deliver safer code, but also deliver faster code, and deliver code faster. <a href="https://aws.amazon.com/blogs/security/an-unexpected-discovery-automated-reasoning-often-makes-systems-more-efficient-and-easier-to-maintain/">https://aws.amazon.com/blogs/security/an-unexpected-discovery-automated-reasoning-often-makes-systems-more-efficient-and-easier-to-maintain/</a> 2024-10-21T15:53:32Z https://yabu.me/nevent1qqsdafzz9ud93r67mfv5cn82vjnd9yxkmat9awtngfsqd6ak64kuvyczyp00as6x9x23yjcp9nsymmfycy0u09ck2frnmpje758s6rpqfp9s7dka3ru More great work from Kyle, and another reminder that retrying isn&#39;t the safe best-practice that many folks assume it is: <a href="https://jepsen.io/analyses/jetcd-0.8.2">https://jepsen.io/analyses/jetcd-0.8.2</a> 2024-08-08T15:05:07Z https://yabu.me/nevent1qqswh64xprdrzeufpt8uqgkwc2ltl32x4ldzplmtexsrqh8a77uvqtczyp00as6x9x23yjcp9nsymmfycy0u09ck2frnmpje758s6rpqfp9s7ttnx37 In reply to <a href='/nevent1qqs276c3pmmrvcvhcz9hsg3k0fj97vxh3ukkjxctwaa0lwpylew3q7qs0v3rl'>nevent1q…v3rl</a><br/>_________________________<br/><br/>The tool I use most often is P, with a bit of TLA+, Kani, and (lately) Lean. I&#39;ve also been learning more about Dafny, but I (personally) haven&#39;t done anything real with it yet. 2024-08-07T23:43:16Z https://yabu.me/nevent1qqsww9hpeh72gmda58pephe3jrmrx9t0m7hpwu0qv92c08ddzrk3nuczyp00as6x9x23yjcp9nsymmfycy0u09ck2frnmpje758s6rpqfp9s70t5vwp In reply to <a href='/nevent1qqsf0806f9tlgh7vlynmphvm0k0ejguu5zkev8hj5m5sdhzc9l2572s0hhdqw'>nevent1q…hdqw</a><br/>_________________________<br/><br/>I&#39;ve been learning Lean recently to add to my formal toolkit, and it&#39;s been a really eye-opening experience. I don&#39;t have much of a math background, but have already found Lean lets me do things that I found really hard before.<br/> <img src="https://fediscience.org/system/media_attachments/files/112/922/466/896/012/670/original/b2a2bf13d5ba76a5.png"> <br/> 2024-08-07T19:48:25Z https://yabu.me/nevent1qqsf0806f9tlgh7vlynmphvm0k0ejguu5zkev8hj5m5sdhzc9l2572szyp00as6x9x23yjcp9nsymmfycy0u09ck2frnmpje758s6rpqfp9s7u0p0dj Loved this talk by Terence Tao on math, AI, and proof assistants. I suspect we&#39;re going to see a similar effect in software: AI will allow formal specs to be developed faster than traditional programs, and for programs to be generated from those specs. <a href="https://www.youtube.com/watch?v=_sTDSO74D8Q">https://www.youtube.com/watch?v=_sTDSO74D8Q</a><br/> <img src="https://fediscience.org/system/media_attachments/files/112/922/460/279/079/365/original/2adc211215e17f38.png"> <br/> 2024-08-07T19:47:44Z https://yabu.me/nevent1qqsp4x2rkfrgucjrk35ej5xz9w2cy0d5djvk4yr3ala6tp63etlxzhqzyp00as6x9x23yjcp9nsymmfycy0u09ck2frnmpje758s6rpqfp9s7zyqqqg New blog post, on the &#34;you don&#39;t need distributed systems&#34; meme, and the idea that distributed systems are only about scale: <a href="https://brooker.co.za/blog/2024/06/04/scale.html">https://brooker.co.za/blog/2024/06/04/scale.html</a> <br/><br/>Short version: scale is only a tiny part of what makes distributed systems interesting and useful. 2024-06-04T21:47:05Z https://yabu.me/nevent1qqsprlehtp40gl442p6yz73x7kc509yetev3kn7xpwpp255x56ugv7qzyp00as6x9x23yjcp9nsymmfycy0u09ck2frnmpje758s6rpqfp9s73sxz8s The internet is talking about retries and backoff! As we&#39;ve seen over the last day or so, simply retrying without backoff leads to an unbounded increase in work. This, in turn, tends to make overloaded systems even more overloaded. What can we do instead? 2023-07-02T15:59:08Z