HNHacker News
TopNewBestAskShowJobs

evanelias

3,028 karma · joined June 6, 2012

NYC-based software engineer. Founder of skeema.io. Formerly a member of Facebook's db automation team and engineer #4 at Tumblr.

Email: evan @ skeema.io

submissionscomments
evanelias··on WA income tax clears House after 24-hour debate
My small NJ town has its own school district, containing a single K-8 school. Yet it has both a superintendent and a principal.

Grades 9-12 feed into a regional high school, which also has its own school district containing just that school. It also has its own superintendent and principal.

I don't think "redundant" is a strong word for this situation.

evanelias··on WA income tax clears House after 24-hour debate
I don't understand your point or why you are sure of that? New Jersey consistently ranks as having the highest average property tax rate of any state, and it also has relatively high income tax rates.

This is mainly caused by having a ridiculous number of tiny towns and tiny school districts, each with redundant services and employees.

evanelias··on Show HN: AI agents run my one-person company on Gemini's free tier – $0/month
> For anything meaningful — client conversations, community discussions like this one — it's always me.

In a six-minute time period, you posted 10 different comments here, totaling nearly 800 words. I don't believe you are being truthful.

evanelias··on Show HN: AI agents run my one-person company on Gemini's free tier – $0/month
The comment you're replying to only asked a single question, which you actually failed to answer.
evanelias··on The next generations of Bubble Tea, Lip Gloss, and Bubbles are available now
Yeah, similar feelings here. I grew up in the BBS world in the 90s, and love a quality TUI experience, but something about this toolset just gives me the ick. I can't keep any of the naming straight; to me it all reads like "combine Chapstick with Cotton Swab inside of Matcha Latte" and my eyes instantly glaze over.

How does this company make money? Is it all just a ZIRP fever dream?

evanelias··on Relicensing with AI-Assisted Rewrite
Wow, as a former MajorMUD addict (~30 years ago) that's extremely interesting to see. Especially since MajorMUD is rarely discussed on HN, even in MUD or BBS-related threads.

Did you find it worked reasonably well on any portion of the codebase you could throw at it? For example, if I recall correctly, all of MajorMUD's data file interactions used the embedded Btrieve library which was popular at the time. For that type of specialized low-level library, I'm curious how much effort it would take to get readable code.

evanelias··on MySQL foreign key cascade operations finally hit the binary log
> The UTF-8 standard was adopted really in 1998-ish and the standard was already variable using 1 to 4 bytes.

No, it was 1 to 6 bytes until RFC 3629 (Nov 2003). AFAIK development of MySQL 4.1 began prior to that, despite the release not happening until afterwards.

Again, they absolutely should have addressed it sooner. But people make mistakes, especially as we're talking about a venture-funded startup in the years right after the dot-com crash.

> It literally does not take any more storage because it is a variable width encoding.

I already addressed that in my previous comment: in old versions of MySQL, a number of critical code paths required allocating worst-case buffer sizes, or accounting for worst-case value lengths in indexes, etc. So if a charset allows 6 bytes per character, that means multiplying max length by 6, in order to handle the pathological case.

evanelias··on Rob Grant, creator of Red Dwarf, has died
Ditto here, season 9 ("Back to Earth") is the only one I have no desire to rewatch.

Season 12 is particularly good though. In my opinion, the first and last episodes of that season are among the funniest they've ever done!

evanelias··on Rob Grant, creator of Red Dwarf, has died
Yeah the first two novels were credited to their "Grant Naylor" partnership, and they're both excellent.

After that, they each wrote an additional Red Dwarf novel individually / separately. Personally I've never come across those last two novels, although I always check for them whenever visiting a used book store. Maybe they were only released in the UK. They're available on Amazon in the US, but I haven't quite given up hope on stumbling across them naturally yet...

evanelias··on Halt and Catch Fire: TV’s best drama you’ve probably never heard of (2021)
> So many quotable moments. Vantasner Danger Meridian. Structural Dynamics of Flow. The attaché badge.

My friends and I have found that Patriot phrases make excellent team names in pub trivia, even if no other teams get the references. (Also "double great", "disgraced veterinarians", "pin-flam-fastened pan traps", more I'm forgetting... clearly I need to rewatch the show yet again!)

> I’m not surprised it didn’t get more traction with the general public

I suspect the show would have had a larger audience if it had a better title. That seems to be an instant negative whenever I try recommending it to someone. Hopefully Steven Conrad's new hbo show will raise his profile enough to let more people rediscover Patriot. It's an absolute gem.

evanelias··on World's Only Pencil Sharpener Museum
Ha, I was just about to comment the same! It's strangely compelling and very memorable. And all of Peaks Island is quite scenic, makes a great summer day trip from Portland Maine.
evanelias··on Databases should contain their own Metadata – Use SQL Everywhere
> something that needs to run on top of a Petabytes of data in the cloud, on Iceberg, with an advanced query planner and a high speed SIMD engine

Part of my confusion was that this blog post makes no mention whatsoever of any of those things!

It gave me the (incorrect) impression that this observability functionality was the purpose of the product. And it is worded in a way which makes no mention of prior art in built-in DBMS observability.

Looking at the other threads here, I don't think I'm the only one who was confused about that. A couple intro paragraphs to the product might help a lot.

evanelias··on Databases should contain their own Metadata – Use SQL Everywhere
This sounds conceptually similar to performance_schema [1] in MySQL or MariaDB, which is a built-in feature originally introduced in MySQL 5.5 (2010). Or perhaps the easier-to-use sys schema [2], which wraps performance_schema among other things, introduced in MySQL 5.7 (2015).

It's great to have that observability functionality, but I don't really understand the purpose of writing a new DBMS from scratch to add this though. Why not get something merged into Postgres core?

[1] https://dev.mysql.com/doc/refman/8.4/en/performance-schema.h...

[2] https://dev.mysql.com/doc/refman/8.4/en/sys-schema.html

evanelias··on MySQL foreign key cascade operations finally hit the binary log
> foreign key cascades where not in the binlogs.

True, but that's finally solved now.

> I also think that some changes in the database layout could lead to strange things on the replicas.

I've been using MySQL for 23 years and have no idea what you're referring to here, sorry. But it's not like other DBs have quirk-free replication either. Postgres logical replication doesn't handle DDL at all, for example.

evanelias··on MySQL foreign key cascade operations finally hit the binary log
InnoDB became the default storage engine over 15 years ago. MyISAM is barely used by anyone today.

What "strange stuff around replication" are you referring to?

evanelias··on MySQL foreign key cascade operations finally hit the binary log
You can tune binlog_expire_logs_seconds to control how long old binlog files stay around. The default is 2592000 seconds (30 days) which is often too long.
evanelias··on MySQL foreign key cascade operations finally hit the binary log
The sync_binlog server variable controls this behavior. The default of 1 means to fsync every time, which is best for durability but worst for performance. See https://dev.mysql.com/doc/refman/8.4/en/replication-options-...
evanelias··on MySQL foreign key cascade operations finally hit the binary log
OK, but that makes perfect sense given utf-16 was actually quite widespread in 2003! For example, Windows APIs, MS SQL Server, JavaScript (off the top of my head)... these all still primarily use utf-16 today even. And MySQL also supports utf-16 among many other charsets.

There wasn't a clear winner in utf-8 at the time, especially given its 6-byte-max representation back then. Memory and storage were a lot more limited.

And yes while 6 bytes was the maximum, a bunch of critical paths (e.g. sorting logic) in old MySQL required allocating a worst-case buffer size, so this would have been prohibitively expensive.

evanelias··on MySQL foreign key cascade operations finally hit the binary log
The performance impact depends substantially on whether you've configured it to fsync the binlog on every group commit.

Also, it's important to consider that replication and backups serve different purposes. Backups alone are insufficient for high availability, change data capture, point-in-time recovery / undoing a bad change, etc.

evanelias··on MySQL foreign key cascade operations finally hit the binary log
They should have addressed it much earlier, but it makes way more sense in historical context: when MySQL added utf8 support in early 2003, the utf8 standard originally permitted up to 6 bytes per char at that time. This had excessive storage implications, and emoji weren't in widespread use at all at the time. 3 bytes were sufficient to store the majority of chars in use at that time, so that's what they went with.

And once they made that choice, there was no easy fix that was also backwards-compatible. MySQL avoids breaking binary data compatibility across upgrades: aside from a few special cases like fractional time support, an upgrade doesn't require rebuilding any of your tables.

evanelias··on An AI agent published a hit piece on me
That's an absolutely ridiculous assertion. Do you similarly think that the Battlestar Galactica reboot was a thinly-veiled racist show because they frequently called the Cylons "toasters"?
evanelias··on Reports of Telnet's death have been greatly exaggerated
Well, that certainly explains why no one in the US telnet BBS community seemed to be discussing having connectivity problems.
evanelias··on Sqldef: Idempotent schema management tool for MySQL, PostgreSQL, SQLite
Right, but my point was that renames in particular typically can't go out well before the corresponding application change [1]. Thus, renames are "out of band" relative to the company's normal schema change process. (This is orthogonal to how schema changes are always "out of band" relative to code deploys; that wasn't what I was referring to.)

[1] In theory a custom ORM could have some kind of dynamic conditional logic for table or column renames, i.e. some way to configure it to retry a query with the "new" name if the query using the "old" name fails. But that has a huge perf impact, and I'm not aware of any common ORMs that do this. So generally if you want to rename a table or column that is already used in prod, there's no way to do it without causing user-facing errors or having system downtime during the period between the SQL rename DDL and the application code change redeploy.

evanelias··on Sqldef: Idempotent schema management tool for MySQL, PostgreSQL, SQLite
pgroll is definitely interesting; it overlaps between both schema management tools and online schema change tools. At scale, I would want to have a better understanding on the performance implications of putting views in front of hot tables. I don't like its use of YAML. iiuc, it is an imperative migration tool, not a declarative tool.

Grate is also imperative, not declarative.

That last link (david.rothlis.net) is about a declarative tool for sqlite, not Postgres.

evanelias··on Postgres Postmaster does not scale
> I don't understand how you can run multiple postmasters.

I believe they're just referring to having several completely-independent postgres instances on the same host.

In other words: say that postgres is maxing out at 2000 conns/sec. If the bottleneck actually was fork rate on the host, then having 2 independent copies of postgres on a host wouldn't improve the total number of connections per second that could be handled: each instance would max out at ~1000 conns/sec, since they're competing for process-spawning. But in reality that isn't the case, indicating that the fork rate isn't the bottleneck.

evanelias··on No More Hidden Changes: How MySQL 9.6 Transforms Foreign Key Management
The major players tried teaming up 12 years ago but it didn't work out: https://en.wikipedia.org/wiki/WebScaleSQL

The general response from the community often included complaints of there being too much fragmentation / too many forks, and that situation has only gotten worse since then.

Also MariaDB looks increasingly well-positioned to gain marketshare. They're making some strategic steps to improve MySQL compat and ease-of-migration.

evanelias··on Sqldef: Idempotent schema management tool for MySQL, PostgreSQL, SQLite
Not sure what that means, but it's named partially as a nod to Skee-Ball. The town I grew up in was actually the home of the factory where all Skee-Ball machines were made.

I was using a location-related naming scheme in general at that time; similarly my automation library was called Go La Tengo because I was living in the town where the band Yo La Tengo was from.

evanelias··on Sqldef: Idempotent schema management tool for MySQL, PostgreSQL, SQLite
For both data migrations and renames, there isn't really a one-size-fits-all solution. That's actually true when doing data changes or renames in imperative (incremental) migrations tools too; they just don't acknowledge it, but at scale these operations aren't really viable. They inherently involve careful coordination alongside application deploys, which cannot be timed to occur at the exact same moment as the migration completion, and you need to prevent risk of user-facing errors or data corruption from intermediate/inconsistent state.

With row data migrations on large tables, there's also risk of long/slow transactions destroying prod DB performance due to MVCC impact (pile-up of old row versions). So at minimum you need to break up a large data change into smaller chunked transactions, and have application logic to account for these migrations being ongoing in the background in a non-atomic fashion.

That all said, to answer from a mechanical standpoint of "how do companies using declarative schema management also handle data migrations or renames":

At large scale, companies tend to implement custom/in-house data migration frameworks. Or for renames, they're often just outright banned, at least for any table with user-facing impact.

At smaller scale, yeah you can just pair a declarative tool for schema changes with an imperative migration tool for non-schema changes. They aren't really mutually exclusive. Some larger schema management systems handle both / multiple paradigms.

For MySQL/MariaDB with Skeema in particular, a few smaller-scale data migration approaches are discussed in a separate post, https://www.skeema.io/blog/2024/07/23/data-migrations-impera...

evanelias··on Sqldef: Idempotent schema management tool for MySQL, PostgreSQL, SQLite
Thank you for the kind words about Skeema!
evanelias··on Sqldef: Idempotent schema management tool for MySQL, PostgreSQL, SQLite
Very large tech companies completely automate the schema change process (at least for all common operations) so that development teams can make schema changes at scale without direct DBA involvement. The more sophisticated companies handle this regardless of table size, sharding, operational events, etc. It makes a massive difference in execution speed for the entire company.

Renames aren't compatible with that automation flow though, which is what I meant by "out-of-band". They rely on careful orchestration alongside code change deploys, which gets especially nasty when you have thousands of application servers and thousands of database shards. In some DBMS, companies automate them using a careful dance of view-swapping, but that seems brittle performance-wise / operationally.

← PreviousPage 4 of 34Next →