Fuzzer Blind Spots (Meet Jepsen!) – https://tigerbeetle.com/blog/2025-06-06-fuzzer-blind-spots-m...
This report found a number of issues, which might be a cause for concern. But I think it's a positive because they didn't just fix the issues, they've expanded their internal test suite to catch similar bugs in future. With such an approach to engineering I feel like in 10 years TigerBeetle would have achieved the "just use Postgres" level of default database in its niche of financial applications.
Also great work aphyr! I feel like I learned a lot reading this report.
Yes, we have around 6,000+ assertions in TigerBeetle. A few of these were overtight, hence some of the crashes. But those were the assertions doing their job, alerting us that we needed to adjust our mental model, which we did.
Otherwise, apart from a small correctness bug in an internal testing feature we added (only in our Java client and only for Jepsen to facilitate the audit) there was only one correctness bug found by Jepsen, and it didn’t affect durability. We’ve written about it here: https://tigerbeetle.com/blog/2025-06-06-fuzzer-blind-spots-m...
Finally, to be fair, TigerBeetle can (and is tested) to survive more faults than Postgres can, since it was designed with an explicit storage fault model and using research that was not available at the time when Postgres was released in ‘96. TB’s fault models are further tested with Deterministic Simulation Testing and we use techniques such as static memory allocation following NASA’s Power of Ten Rules for Safety-Critical Code. There are known scenarios in the literature that will cause Postgres to lose data, which TigerBeetle can detect and recover from.
For more on this, see the section in Kyle’s report on helical fault injection (most Raft and Paxos implementations were not designed to survive this) as well as a talk we gave at QCon London: https://m.youtube.com/watch?v=_jfOk4L7CiY
I have followed TigerBeetle with interest for a while, and thank you for your inspirational work and informative presentations.
However, you have stated in several occasions that the lack of memory safety in Zig is not a concern since you don't dynamically allocate memory post startup. However, one of the defects uncovered here (#2435) was caused by dereferencing an uninitialized pointer. I find this pretty concerning, so I wonder if there is something that you will be doing differently to eliminate all similar bugs going forward?
https://csrc.nist.gov/projects/automated-combinatorial-testi...
My intro to other categories of test generation was usually this paper:
https://cs.stanford.edu/people/saswat/research/ASTJSS.pdf
Maybe see of your team can build combinatorial- or path-based testing in Zig next.
Are they enabled in production? Are there some expensive ones that aren’t?
What are you referencing here?
Question about how people end up using TigerBeetle. There's presumably a lot of external systems and other databases around a TigerBeetle install for everything that isn't an Account or Transfer. What's the typical pattern for those less reliable systems to square up to TigerBeetle, especially to recover from consistency issues between the two?
The typical pattern in integrating TigerBeetle is to differentiate between control plane (Postgres for general purpose or OLGP) and data plane (TigerBeetle for transaction processing or OLTP).
All your users (names, addresses, passwords etc.) and products (descriptions, prices etc.) then go into OLGP as your "filing cabinet".
And then all the Black Friday transactions these users (or entities) make, to move products from inventory accounts to shopping cart accounts, and from there to checkout and delivery accounts—all these go into OLTP as your "bank vault". TigerBeetle lets you store up to 3 user data identifiers per account or transfer to link events (between entitites) back to your OLGP database which describes these entities.
This architecture [1] gives you a clean "separation of concerns", allowing you to scale and manage the different workloads independently. For example, if you're a bank, it's probably a good idea not to keep all your cash in the filing cabinet with the customer records, but rather to keep the cash in the bank vault, since the information has different performance/compliance/retention characteristics.
This pattern makes sense because users change their name or email address (OLGP) far less frequently than they transact (OLTP).
Finally, to preserve consistency, on the write path, you treat TigerBeetle as the OLTP data plane as your "system of record". When a "move to shopping cart" or "checkout" transaction comes in, you first write all your data dependencies to OLGP if any (and say S3 if you have related blob data) and then finally you commit your transaction by writing to TigerBeetle. On the read path, you query your system of record first, preserving strict serializability.
Does that make sense? Let me know if there's anything here we can drill into further!
[1] https://docs.tigerbeetle.com/coding/system-architecture/
It looks like the segfaults on the JNI side would not have been protected if Rust or some other memory safe language were being used - the lack of memory safety bugs gives some decent proof that TigerBeetle's approach to Zig programming (TigerStyle iirc, lol) does what it sets out to do.
EDIT: But, yeah, totally, if not for TigerStyle, we'd die to nasal demons!
Special kudos to the founders who are sharing great insights in this thread.
Appreciate your kind words too, and look forward also to sharing something new in our talks at SD25 in Amsterdam soon!
This is an interesting meta pattern where doing something _harder_ actually simplifies the system.
Another example is that, because we assume that the disk can fail and need to include repair protocol, we get state-synchronization for a lagging replica "for free", because it is precisely the same situation as when the entire disk gets corrupted!
In TigerBeetle, we take advantage of some special properties to make the state machine checking part linear-time. We let TigerBeetle tell us exactly which transactions happen. We can do this because it's a.) strong serializable, b.) immutable (in that we can inspect DB state to determine whether an op took place), and c.) exposes a totally ordered timestamp for every operation. Then we check that that timestamp order is consistent with real-time order, using a linear-time cycle detection approach called Elle. Having established that TigerBeetle's claims about the timestamp order are valid, we can apply those operations to a simulated version of the state machine to check semantic correctness!
I'd like to generalize this to other systems, but it's surprisingly tricky to find all three of those properties in one database. Maybe an avenue for future research!
I think it should be http://pmg.csail.mit.edu/papers/vr-revisited.pdf (http scheme not https) ?
And now I have some Friday evening reading material.
The VSR 2012 paper is one of my favorites as is “Protocol-Aware Recovery for Consensus-Based Storage”, which is so powerful.
Hope you enjoy the read!
After reading the Jepsen report on TigerBeetle, the related blog post, and briefly reviewing the Antithesis integration code on GitHub workflow, I'm trying to better understand the testing scope.
My core question is: could these bugs detected by the Jepsen test suite have also been found by the Antithesis integration?
This question comes from a few assumptions I made, which may be incorrect:
- I thought TigerBeetle was already comprehensively tested by its internal test suite and the Antithesis product.
- I had the impression that the Antithesis test suite was more robust than Jepsen's, so I was surprised that Jepsen found an issue that Antithesis apparently did not.
I'm wondering if my understanding is flawed. For instance:
1. Was the Antithesis test suite not fully capable of detecting this specific class of bug?
2. Was this particular part of the system not yet covered by the Antithesis tests?
3. Am I fundamentally comparing apples and oranges, misunderstanding the different strengths and goals of the Jepsen and Antithesis testing suites?
I would greatly appreciate any insights that could help me understand this better. I want to be clear that my goal is to educate myself on these topics, not to make incorrect assumptions or assign responsibility.
There are almost certainly blind spots in the Jepsen test generators too--that's part of why designing different generators is so helpful!
1. Some sort of environment, which can run the system. The simplest environment is to spin up a real cluster of machines, but ideally you want something fancier, to improve performance, control over responses of external APIs, determinism, reproducibility, etc. 2. Some sort of load generator, which makes the system in the environment do interesting thing 3. Some sort of auditor, which observes the behavior of the system under load and decides whether the system behaves according to the specification.
Antithesis mostly tackles problem #1, providing a deterministic simulation environment as a virtual machine. The same problem is talked by jepsen (by using real machines, but injecting faults at the OS level), and by TigerBeetle's own VOPR (which is co-designed with the database, and for that reason can run the whole cluster on just a single thread). There there approaches are complimentary and are good at different things.
For this bug, the critical part was #2, #3 --- writing workload verifier and auditor that actually can trigger the bug. Here, it was aphyr's 1600 lines of TigerBeetle-specfic Clojure code that triggred and detected the bug (and then we patched _our_ equivalent to also trigger it. Really, what's buggy here is not the database, but the VOPR. Database having bugs is par of course, you can't just avoid bugs through the sheer force of will. So you need testing strategy that can trigger most bugs, and any bug that slips through is pointing to the deficiency in the workload generator.)
https://github.com/jepsen-io/tigerbeetle/blob/main/src/jepse...
To understand why the query engine bug slipped through, see: https://tigerbeetle.com/blog/2025-06-06-fuzzer-blind-spots-m...
At a national level, we’re working with the Gates Foundation to integrate TigerBeetle into their non-profit central bank switch that will be powering Rwanda’s National Digital Payments System 2.0 later this year [1].
At an enterprise level, TigerBeetle already powers customers processing 100M+ transactions per month in production, and we recently signed our first $2B fintech unicorn in Europe with a few more in the US about to close. Because of the move to realtime transaction processing around the world [2] there’s been quite a bit of interest from companies wanting to move to TigerBeetle for more performance.
Finally, to your question, some of the founders of Clear Street, a fairly large brokerage on Wall Street have since invested [3] in TigerBeetle.
[1] https://mojaloop.io/how-mojaloop-enables-rndps-2-0-ekash/
[2] https://tigerbeetle.com/blog/2024-07-23-rediscovering-transa...
"Invested" in terms of "giving you money" or in terms of "Now uses the database themselves"? I read it as the first, but I think the question is about usage, not investments.
We don't publish the pre-compiled libs yet, but it will be available soon when we stabilize the API. For now, it can be built locally from source.
Example of using the C client: https://github.com/tigerbeetle/tigerbeetle/blob/main/src/cli...
Or possibly my mental model of how physical disks and the driver stack behave these days is outdated.
For more on our open source thinking and how this is orthogonal to business model (and product!), see our interview with the Changelog: https://m.youtube.com/watch?v=Yr8Y2EYnxJs
1. AWS will take your initial and ongoing investment in the implementation but they don't have to share theirs with you. Specifically, they will take your improvements but their own improvements (say some performance optimizations) they can keep to themselves. It's good business sense if it allows them to further differentiate their "improved" offering from your "vanilla" service.
2. Competing on the the interface in this case really means competing on related services like management, etc. So your thesis is that you will provide a better/cheaper managed service than AWS. Even if that's true (a big if), most of the time the decision which service to use will have little to do with technical merit. I.e. we already use AWS, have SLA painfully negotiated, get volume discounts, etc. Do we really want to go through all of this with another vendor just for one extra service.
Just a couple of thoughts that will hopefully help you sharpen your thesis.
That's like saying that rice noodles are no good for making risotto. At the core they are both rice...
Yes, TigerBeetle specializes only for transaction processing (OLTP). It’s not a general-purpose (OLGP) DBMS.
That said, we have customers from energy to gaming, and of course fintech.
