Rendered at 14:26:11 GMT+0000 (Coordinated Universal Time) with Cloudflare Workers.
dangoodmanUT 11 hours ago [-]
> DbOptions::durable()
> Appended to the WAL without a per-write sync
So… it’s not durable? Durable doesn’t mean “survives a process restart”, it means “durably saved to persistent storage”. For example, this “durable” mode wouldn’t survive power loss.
stingraycharles 10 hours ago [-]
Yeah this should be benchmarked against other systems that have flush() disabled.
mmap is nice but it doesn’t support durable semantics in the way that we usually mean with databases.
if a write is acknowledged it should not be forgotten, which is not what this is.
muvlon 2 hours ago [-]
As of a couple years ago, mmap actually has a MAP_SYNC flag that makes it durable in the DB sense. The caveat is that it requires DAX on the file and so comes with a whole bunch of restrictions w.r.t. filesystem, storage media and even CPU architecture.
rgbimbochamp 10 hours ago [-]
You're right, that mode provides process crash recovery, not power-loss durability. The benchmark compares it against fjall’s equivalent buffered-WAL mode.
deathanatos 7 hours ago [-]
Word choice matters. Defaults matter. People will go "well it says durable right here" and while arguably, yes, they should RTFM, it would still be great if tool-builders did not set the shotgun's default state to State::AT_FOOT. It would be nice if every paragraph of technical writing that I have to do need not be burdened by a thousand asterisks of "durable in this context means something other than durable".
carlmr 6 hours ago [-]
>arguably, yes, they should RTFM
Agreed. Good design is when the things do what you expect them to do without reading the manual, don't reuse wording with other meaning in the wrong way. That way if you do encounter nee wording, you know you should read the manual.
If that's your design constraint, couldn't you speed it up by getting rid of the WAL?
oneshadab 8 hours ago [-]
You'd lose durability against process crashes.
If your system has a reasonable tolerance for power failure (multi-az multi-cloud), this can provide much better throughput
t098i3 8 hours ago [-]
Indeed, a common enough pattern for etcd is to run it backed by a RAMdisk and have multi-az availability + periodic backups + tolerance at a business level to be OK losing some recent data.
procaryote 7 hours ago [-]
Pretty much... paranoid() seems to be the real durable() which isn't a great look for a database project.
Being able to recover a db without corruption beyound losing the last few writes is a pretty useful feature, and buys a lot of performance, but it would be better to label that clearly, as a reasonable expectation on the durable() preset would be for it to be Durable.
deepsun 8 hours ago [-]
Are we back to MongoDB -- no fsync() but webscale speed?
This is surprisingly common, from what I can tell.
KAdot 39 minutes ago [-]
The benchmark is setup to test 80 MiB dataset on a machine with 32 GiB RAM, which doesn't represent a typical database workload. How does the key-value store perform on larger than RAM datasets? MMAP is fast when the dataset fits in memory, but it can slow to a crawl when it doesn't, especially if the workload is mostly random point lookups.
boguscoder 10 hours ago [-]
Embedded could also mean no_std, which this is absolutely not. Still cool though
LoganDark 9 hours ago [-]
Yep, the correct term here is "embeddable", not embedded.
mattrighetti 5 hours ago [-]
I like the fact that the first commits were about the logo, important things first :D
And it's not complete until it can send and receive email.
idontneedcoffee 2 hours ago [-]
No way(hides)
otabdeveloper4 8 hours ago [-]
A file-backed hashtable isn't really a DB.
rcxdude 7 minutes ago [-]
it's not a relational database, but those aren't the only kind.
insanitybit 4 hours ago [-]
If someone asked me what a DB is, I'd probably start the conversation with the exact description "a file-backed map, like a hashmap".
riffraff 8 hours ago [-]
Well, dbm (database manager) is basically that and has been called that for almost 50 years
otabdeveloper4 36 minutes ago [-]
dbm isn't a database either, and people who call it that are simply wrong.
mikedelago 3 hours ago [-]
Why not?
dist-epoch 4 hours ago [-]
How does it compare to RocksDB?
DarmokTanagra 4 hours ago [-]
Off topic, but why is tokio still independent of the rust async runtime?
It seems pretty ubiquitous yet not a part of the core rust libs.
kshri24 2 hours ago [-]
I prefer compio over tokio. So would not want libs to depend on a library that is essentially a runtime. Runtimes should be swappable.
insanitybit 4 hours ago [-]
The same reason as ever. Not everyone wants to use the same runtime.
DarmokTanagra 4 hours ago [-]
Sure, but now if my rust application isn't using tokio I have to include it as a new dependency because the author of this lib decided to use it as his async runtime?
I'm not trying to be pedantic, but this split over async runtimes was what originally turned me off of rust years ago and it still seems to be an issue.
timschmidt 3 hours ago [-]
I'm using async on esp32 with embassy, for instance, where there's simply no room and no need for a runtime as complex as tokio.
airstrike 2 hours ago [-]
The author of this library can leave the async library choice up to the user.
iced does that quite well.
insanitybit 3 hours ago [-]
You would only have to include it if the library uses `spawn`, as far as I am aware, or some tokio specific type, which is the same as any other library.
tinco 3 hours ago [-]
It's not pedantry, you're trying to find something in Rust that's simply not there. If this sort of thing turns you off on Rust you should be looking at a programming language with other priorities.
derriz 3 hours ago [-]
I’ve had the same reaction. We’ve seen this play out in other language eco-systems (Java - JAXP, javax.validation, JPA, etc all ended up with a de facto single implementation) and the idea of pluggable implementations sounds appealing but rarely pays off. The price that Rust paid for this abstraction - which in fact ended up not being useful as tokio is the only reasonable choice - was high in terms of requiring ugly (to my eyes) changes to its type system.
VorpalWay 3 hours ago [-]
Tokio is definitely not the only reasonable choice. I use embassy[1] on microcontrollers (the Pi Pico and ESP32 in my case), where tokio can't run.
There is no price to pay on the type system that was imposed by tokio. I assume you're saying something like "I have to add 'static in generics" or something?
It has absolutely paid off, there are many people not using tokio.
Tepix 7 hours ago [-]
Is an atomic get+delete operation planned?
Surac 5 hours ago [-]
Fast compared to what?
nine_k 11 hours ago [-]
I suppose the insane speed is due to this:
> TurboKV's persisted Bloom-filter format uses hardware AES.
Also, built-in LZ4 compression.
I would expect SIMD to be used for scans.
haberman 11 hours ago [-]
I assume this is for hashing. I've seen several hashing algorithms turn to hardware AES instructions before, but I haven't seen any evidence that this technique outperforms state-of-the-art hashes like RapidHash (https://github.com/Nicoshev/rapidhash) in either quality or speed.
itemize123 3 hours ago [-]
that's optimizing a pretty fast already portion of code. unlikely it's the difference maker.
Sesse__ 6 hours ago [-]
For any complex system, there's never one single trick or design choice that makes it fast. It's always a large amount of engineering (or exaggerations, of course).
rgbimbochamp 11 hours ago [-]
Those help but the main write speed gain is the WAL, that uses preallocated mmap segments to avoid a write(2) per durable mutation while preserving crash recovery.
AES hashing mainly helps Bloom filter point lookups and LZ4 mainly helps SSTable I/O. Scans benefit indirectly, but don’t yet use a custom SIMD merge loop.
bestouff 10 hours ago [-]
I said elsewhere this doesn't survive a power loss.
rgbimbochamp 8 hours ago [-]
paranoid() does survive power loss.
taneq 10 hours ago [-]
While it’s important to make this explicit, at what point do we just assume a high-reliability UPS is table stakes?
Of course, if you need SIL2 type reliability then you need to assume any given hardware component can spontaneously combust and become a total loss, at which point the data loss caused by a power cut is a rounding error.
toast0 8 hours ago [-]
> While it’s important to make this explicit, at what point do we just assume a high-reliability UPS is table stakes?
Several years after they become commercially available?
My experience with small UPSes is they tend to cook the batteries and you don't find out until they switch the load and the battery doesn't hold up.
Large facility UPSes tend to do better, but automatic transfer switches have a tendancy to fail ocassionally. If you're hosted in many locations, it's not unusual to have a couple ATS failures per decade.
All that said, unexpected power loss is certainly one reason that writes may be lost, but OSes crash too. Disk firmware can also crash, but if thst bricks the disk, writes in progress don't really matter. Sometimes cabling fails. Or you get a uncorrectable ECC error (which will typically cause an OS panic... unless you're running a very fancy OS, but if it's in dirty disk backed page, even a fancy OS wouldn't save you)
Plenty of applications don't need or want to pay the cost for full commit to disk, but calling something durable when it's not committed to disk is inaccurate.
And that's before we get into the whole thing where the OS and the disk like to return success when things haven't quite finished.
whilenot-dev 9 hours ago [-]
What's got this to do with a UPS? Not having a UPS is an external threat on the reliability of the power grid.
Doing a hard shutdown or tripping over power cords seem much likelier local scenarios than any spontaneous combustion of hardware components.
zbentley 2 hours ago [-]
A UPS won’t save you from a kernel panic.
imtringued 5 hours ago [-]
Not sure how you managed to do it but you got it completely backwards.
If someone demands that the database should use fsync and only respond with success once the write finished, it is not some arbitrarily high reliability demand that needs to be implemented using reliable hardware. In fact, the entire point of implementing the power loss protection in software is so that you don't need perfectly reliable hardware. The power loss event turns into a downtime event which is often completely acceptable.
The requirement to have infallible hardware only emerged because the software refused to do its job. Infallible hardware is not a requirement decided by the user, it's a requirement decided by the developer of TurboKV to intentionally restrict his software to exclusively operate in a reliable hardware environment.
The fact that the user specified durability of the KV store during power loss does not make the user obsessed over hardware reliability, the software shifted the burden onto the hardware and forced the user to deal with this mess.
I don't know how exactly TurboKV works so let's talk about a hypothetical software instead.
Let's say the software cannot survive a power loss event and just corrupts the database. If the user wants to operate the software, he is forced by the software to operate it in an infallible environment where power loss can never occur. Based on how the software was designed, power loss is a catastrophic event. The SIL2 type reliability you're talking about only makes sense in contexts with catastrophic events.
So how it went is that the user made a reasonable demand with bounded reliability: "please survive power loss with durable writes" and the author says, sure just run the software on a SIL2 type reliability hardware environment.
It's not the user who blew up the hardware requirements.
dangoodmanUT 11 hours ago [-]
Iirc that’s how badger handles the WAL as well
9 hours ago [-]
AtlasBarfed 10 hours ago [-]
Aphyr or gtfo
rollulus 7 hours ago [-]
Now that “blazing fast in Rust” has become a meme, is “insanely” the next thing?
> Appended to the WAL without a per-write sync
So… it’s not durable? Durable doesn’t mean “survives a process restart”, it means “durably saved to persistent storage”. For example, this “durable” mode wouldn’t survive power loss.
mmap is nice but it doesn’t support durable semantics in the way that we usually mean with databases.
if a write is acknowledged it should not be forgotten, which is not what this is.
Agreed. Good design is when the things do what you expect them to do without reading the manual, don't reuse wording with other meaning in the wrong way. That way if you do encounter nee wording, you know you should read the manual.
If your system has a reasonable tolerance for power failure (multi-az multi-cloud), this can provide much better throughput
Being able to recover a db without corruption beyound losing the last few writes is a pretty useful feature, and buys a lot of performance, but it would be better to label that clearly, as a reasonable expectation on the durable() preset would be for it to be Durable.
https://github.com/ncruces/go-sqlite3
And I'm still having fun.
https://github.com/ncruces/go-sqlite3/pull/421
It seems pretty ubiquitous yet not a part of the core rust libs.
I'm not trying to be pedantic, but this split over async runtimes was what originally turned me off of rust years ago and it still seems to be an issue.
iced does that quite well.
[1] https://embassy.dev/
It has absolutely paid off, there are many people not using tokio.
> TurboKV's persisted Bloom-filter format uses hardware AES.
Also, built-in LZ4 compression.
I would expect SIMD to be used for scans.
Of course, if you need SIL2 type reliability then you need to assume any given hardware component can spontaneously combust and become a total loss, at which point the data loss caused by a power cut is a rounding error.
Several years after they become commercially available?
My experience with small UPSes is they tend to cook the batteries and you don't find out until they switch the load and the battery doesn't hold up.
Large facility UPSes tend to do better, but automatic transfer switches have a tendancy to fail ocassionally. If you're hosted in many locations, it's not unusual to have a couple ATS failures per decade.
All that said, unexpected power loss is certainly one reason that writes may be lost, but OSes crash too. Disk firmware can also crash, but if thst bricks the disk, writes in progress don't really matter. Sometimes cabling fails. Or you get a uncorrectable ECC error (which will typically cause an OS panic... unless you're running a very fancy OS, but if it's in dirty disk backed page, even a fancy OS wouldn't save you)
Plenty of applications don't need or want to pay the cost for full commit to disk, but calling something durable when it's not committed to disk is inaccurate.
And that's before we get into the whole thing where the OS and the disk like to return success when things haven't quite finished.
Doing a hard shutdown or tripping over power cords seem much likelier local scenarios than any spontaneous combustion of hardware components.
If someone demands that the database should use fsync and only respond with success once the write finished, it is not some arbitrarily high reliability demand that needs to be implemented using reliable hardware. In fact, the entire point of implementing the power loss protection in software is so that you don't need perfectly reliable hardware. The power loss event turns into a downtime event which is often completely acceptable.
The requirement to have infallible hardware only emerged because the software refused to do its job. Infallible hardware is not a requirement decided by the user, it's a requirement decided by the developer of TurboKV to intentionally restrict his software to exclusively operate in a reliable hardware environment.
The fact that the user specified durability of the KV store during power loss does not make the user obsessed over hardware reliability, the software shifted the burden onto the hardware and forced the user to deal with this mess.
I don't know how exactly TurboKV works so let's talk about a hypothetical software instead.
Let's say the software cannot survive a power loss event and just corrupts the database. If the user wants to operate the software, he is forced by the software to operate it in an infallible environment where power loss can never occur. Based on how the software was designed, power loss is a catastrophic event. The SIL2 type reliability you're talking about only makes sense in contexts with catastrophic events.
So how it went is that the user made a reasonable demand with bounded reliability: "please survive power loss with durable writes" and the author says, sure just run the software on a SIL2 type reliability hardware environment.
It's not the user who blew up the hardware requirements.