The comments are pretty interesting. There was a discussion about how different they were trying to make the public image about their tech look compared to their actual usage - supposedly to entice more engineers to join.
I think its interesting to attach any company prominently to a database technology since theoretically there would be varied use cases across an org like uber which would likely want different technologies depending on those use cases. Of course they might just have 50 other articles like this for all the other tech they use.
> The [Odin] platform supports 23 technologies, ranging from traditional online databases such as MySQL® and Cassandra® to advanced data platform technologies, including HDFS™, Presto™, and Kafka®.
That all depends on the setup. The "standard" setup (not specific to MySQL) is:
- Single Write Leader per partition
- Backup Write Leader that is setup with synchronous replication (so WL -> WLB and waits for commit)
- Read Followers all connected asynchronously using either binlog replication (not recommended anymore) or GTID-based row replication (recommended)
In the above scenario, the odds of loss are pretty small since the Write Leader has a direct backup, and any of the Read Followers can be promoted to a Write Leader/Backup. DDIA calls the above semi-synchronous replication, although MySQL now supports a similar-but-slightly different version out of the box: https://dev.mysql.com/doc/refman/8.4/en/replication-semisync...
Every six months, I explore switching from MySQL to something new for a more modern tech stack. However, MyRocks (https://docs.percona.com/percona-server/8.4/myrocks-index.ht...) is truly impressive. It allows me to efficiently compress my text-rich rows.
Read the parent post to this one, AI slop has an uncanny valley associated with it. Somehow it sticks out like a sore thumb but i can't put my finger on why.
I don't know, I didn't spot that this was AI generated. Perhaps because there's some truth that MyRocks is actually really good at compression.
Back when I was exploring migrating from TokuDB to MyRocks the only problem with it was that it didn't have a file per partition, meaning if you were doing retention you couldn't just drop old daily partitions cheaply.
I ran this through ChatGPT and it found a few grammatical errors - that I agree with - and awkward wording. Usually AI generated text doesn’t have these types of errors.
It reads like a standard corporate blog post. You can find plenty of those on AWS blogs that were written before LLMs were publicly available.
Is it just me or all the diagram images are broken? It looks like the lazy-loading is not working. I tried on archive.org and archive.is and same issue: the images didn't load.
Edit: yeah those images have their src attribute in the form of “lh7-rt.googleusercontent.com/docsz/[VERY LONG SLUGS]” and a quick look at the dev console shows that Google returns “429 Too Many Requests” for all of them.
I guess they just copy-pasted the URLs from some Google Doc and tried to hotlink them here? Interestingly if you access these directly (instead of embedded into Uber's blog) then they display fine and Google doesn't complain with a 429.
I’m amazed how I’ve worked with all of the critical technologies in this article, I’ve dealt with the same or similar concerns, and yet the author or authors have written this engineering blog post in such a way as to not convey anything meaningful at all.
One part of it is the constant talk of high level abstract infrastructural pieces, and the other is bad product or concept naming.
Odin, “the controller,” the constant obsession with certain engineering orgs to use words like “plane,” and likewise, “fabric” was used at a previous org I worked for.
I’m sure Uber is doing Real™ Work, but this kind of crap sets off all my wank and bullshit alarms.
It’s just clients talking to servers talking to servers talking to proxies talking to servers talking to databases talking to replicas. Can you please stop with the false high engineering bullshit?
the plane separation is important though, you want to ensure your control messages always go through, like how servers have a dedicated line for management.
Absolutely not true in my experience. MySQL has its share of issues (all DBs do) but it is rock-solid when using the correct engine (InnoDB for most cases, RocksDB for high-throughput writes, Memory for caching). MySQL is very hard to beat for very high-volume OLTP workloads, both reads and writes. Its replication systems were years ahead of other systems (SQL Server, Postgres, SQLite doesn't have replication). DuckDB AFAIK is OLAP and they don't compete in the same space. Every DB system has "the things its good at" and MySQL really shines at very high-volume OLTP spread across partitions.
Disgusting company: Uber takes 70% cut from rider's fare, drivers made only 30% on top of that Uber place those stolen earnings in the driver's tax report, not on them.
You simply cannot take 70% from the revenue coming from an employee-like agent, yet report that to the government as being that person's income. Not only would that be blatant fraud, but depending on the exact percentages and absolute amount, the employee might have to give all their pay to the government to covert the tax, and even owe some more after that.
Uber is not a company worth following for technology. It’s like a bunch of juniors using tech wrong and changing tech as a solution to a non existent problem.
More posts about mysql from Uber - it is interesting to compare state of Mysql in Uber with 2016.
* Why Uber Engineering Switched from Postgres to MySQL (2016) https://news.ycombinator.com/item?id=26283348
* Upgrading Uber's MySQL Fleet https://news.ycombinator.com/item?id=41836748
The comments are pretty interesting. There was a discussion about how different they were trying to make the public image about their tech look compared to their actual usage - supposedly to entice more engineers to join.
I think its interesting to attach any company prominently to a database technology since theoretically there would be varied use cases across an org like uber which would likely want different technologies depending on those use cases. Of course they might just have 50 other articles like this for all the other tech they use.
Another interesting one about their Storage Platform (which includes MySQL): https://www.uber.com/blog/odin-stateful-platform
> The [Odin] platform supports 23 technologies, ranging from traditional online databases such as MySQL® and Cassandra® to advanced data platform technologies, including HDFS™, Presto™, and Kafka®.
Mirror with working images: https://web.archive.org/web/20250130201835/https://www.uber....
So if these clusters are using binlog replication, do they just ignore the possibility of lost writes and inconsistent data after a failover?
That all depends on the setup. The "standard" setup (not specific to MySQL) is:
- Single Write Leader per partition
- Backup Write Leader that is setup with synchronous replication (so WL -> WLB and waits for commit)
- Read Followers all connected asynchronously using either binlog replication (not recommended anymore) or GTID-based row replication (recommended)
In the above scenario, the odds of loss are pretty small since the Write Leader has a direct backup, and any of the Read Followers can be promoted to a Write Leader/Backup. DDIA calls the above semi-synchronous replication, although MySQL now supports a similar-but-slightly different version out of the box: https://dev.mysql.com/doc/refman/8.4/en/replication-semisync...
Correct, you would need to use something like galera clustering or synchronous replication to not lose writes.
Every six months, I explore switching from MySQL to something new for a more modern tech stack. However, MyRocks (https://docs.percona.com/percona-server/8.4/myrocks-index.ht...) is truly impressive. It allows me to efficiently compress my text-rich rows.
Read the parent post to this one, AI slop has an uncanny valley associated with it. Somehow it sticks out like a sore thumb but i can't put my finger on why.
I don't know, I didn't spot that this was AI generated. Perhaps because there's some truth that MyRocks is actually really good at compression.
Back when I was exploring migrating from TokuDB to MyRocks the only problem with it was that it didn't have a file per partition, meaning if you were doing retention you couldn't just drop old daily partitions cheaply.
Page is 404-ing now.
This is ai written garbage
I ran this through ChatGPT and it found a few grammatical errors - that I agree with - and awkward wording. Usually AI generated text doesn’t have these types of errors.
It reads like a standard corporate blog post. You can find plenty of those on AWS blogs that were written before LLMs were publicly available.
Your AI detector seems to be in need of recalibration.
I ran the text through several AI detectors and they all returned 0% AI 100% human.
The cover image is however:
> Cover Photo Attribution: The cover photo was generated using OpenAI ChatGPT Enterprise.
this comment is ai written garbage
When people say MySQL nowadays, do they really mean MySQL, or MariaDB?
IME MySQL. Folks using MariaDB aren't shy about saying so.
Is it just me or all the diagram images are broken? It looks like the lazy-loading is not working. I tried on archive.org and archive.is and same issue: the images didn't load.
Edit: yeah those images have their src attribute in the form of “lh7-rt.googleusercontent.com/docsz/[VERY LONG SLUGS]” and a quick look at the dev console shows that Google returns “429 Too Many Requests” for all of them.
I guess they just copy-pasted the URLs from some Google Doc and tried to hotlink them here? Interestingly if you access these directly (instead of embedded into Uber's blog) then they display fine and Google doesn't complain with a 429.
Looks like images are broken in a lot of posts on this blog
it may be that google is checking the referer header.
Broken for me too.
I’m amazed how I’ve worked with all of the critical technologies in this article, I’ve dealt with the same or similar concerns, and yet the author or authors have written this engineering blog post in such a way as to not convey anything meaningful at all.
One part of it is the constant talk of high level abstract infrastructural pieces, and the other is bad product or concept naming.
Odin, “the controller,” the constant obsession with certain engineering orgs to use words like “plane,” and likewise, “fabric” was used at a previous org I worked for.
I’m sure Uber is doing Real™ Work, but this kind of crap sets off all my wank and bullshit alarms.
It’s just clients talking to servers talking to servers talking to proxies talking to servers talking to databases talking to replicas. Can you please stop with the false high engineering bullshit?
the plane separation is important though, you want to ensure your control messages always go through, like how servers have a dedicated line for management.
No mention of Schemaless? Was that retired?
Uber you guys need to delete this, the quality of this post is absolutely embarrassing.
Too late I already archive.org'd it lol: https://web.archive.org/web/20250130201835/https://www.uber....
Looks like Uber's ditched their expensive designers for AI on their corporate blog's lead images.
It reads like AI slop too. Not so sure why it is getting voted up.
People see uber tech blog and jump straight to upvoting.
I remember when that blog used to be top notch.
Today it just seems odd that anybody is still using MySQL. Postgres? Sure. SQLlite? Hell yeah! DuckDB? Of course. MySQL? Not so much.
Absolutely not true in my experience. MySQL has its share of issues (all DBs do) but it is rock-solid when using the correct engine (InnoDB for most cases, RocksDB for high-throughput writes, Memory for caching). MySQL is very hard to beat for very high-volume OLTP workloads, both reads and writes. Its replication systems were years ahead of other systems (SQL Server, Postgres, SQLite doesn't have replication). DuckDB AFAIK is OLAP and they don't compete in the same space. Every DB system has "the things its good at" and MySQL really shines at very high-volume OLTP spread across partitions.
we are almost all in MySQL. we are also old and 500b
This is an embarrassment and just one more example of how far this enterprise has fallen.
Even as a consumer, I've stopped using Uber due to, of all things "biz 101", incompetent billing.
Disgusting company: Uber takes 70% cut from rider's fare, drivers made only 30% on top of that Uber place those stolen earnings in the driver's tax report, not on them.
That sounds extremely improbable.
You simply cannot take 70% from the revenue coming from an employee-like agent, yet report that to the government as being that person's income. Not only would that be blatant fraud, but depending on the exact percentages and absolute amount, the employee might have to give all their pay to the government to covert the tax, and even owe some more after that.
In what country are they doing this?
[dead]
Even the authors listed at end looks like AI generated.
Uber is not a company worth following for technology. It’s like a bunch of juniors using tech wrong and changing tech as a solution to a non existent problem.
I always got the impression their tech stack was designed to be as complicated as possible so they could hire juniors who thought it looked cool.
They used to produce some neat stuff. Jaeger and Cadence were quite cool.