Given that will presumably replicate over time, and increase the PoP's it exists in, replacing the e

Given that will presumably replicate over time, and increase the PoP's it exists in, replacing the entire database will destroy that replication and you'll start from step 1

JJames Any regular process which involves dropping and recreating millions of rows of a...

ZackOP•4/7/25, 11:17 PM

Something like this would be totally fine in postgres, it's an extremely simple operation that won't impact the database whatsoever

ZZack Given that will presumably replicate over time, and increase the PoP's it exists...

Isaac McFadyen•4/7/25, 11:31 PM

Once it lands, D1 replication is extremely quick so this should still be a reasonable method of replacing the DB

ZackOP•4/7/25, 11:32 PM

I've not seen any info on that tbh

ZackOP•4/7/25, 11:34 PM

I imagine the process of replication is fast, but I expect the process of learning where to replicate too is not necessarily as fast

ZackOP•4/7/25, 11:35 PM

And switching out your DB and dumping all load onto a single node while it replicates out is not going to be great

Isaac McFadyen•4/7/25, 11:38 PM

I can't really say too much at the moment other than the fact that switching out a DB will still be a reasonable thing to do once replication lands

ZackOP•4/7/25, 11:41 PM

We'll see whenever that lands I guess, regardless, dropping a table should not be something that takes out a production database

gruntlord6•4/7/25, 11:41 PM

if its anything like other platforms the writes go to all replicas without any work on your end

gruntlord6•4/7/25, 11:41 PM

and it's basically instant

ZackOP•4/7/25, 11:41 PM

If you replace the entire database, there are no replicas

gruntlord6•4/7/25, 11:41 PM

but like they said can't speak for something we don't have

ZZack We'll see whenever that lands I guess, regardless, dropping a table should not b...

Isaac McFadyen•4/7/25, 11:43 PM

Sure but if you want zero downtime, you duplicate the DB, drop the table, and switch the "connection" (binding)

IIsaac McFadyen Sure but if you want zero downtime, you duplicate the DB, drop the table, and sw...

ZackOP•4/7/25, 11:44 PM

Yes, that is something we've considered, it's just more work

ZackOP•4/7/25, 11:44 PM

We're already having to work around multiple other issues that we didn't expect lol

く

くもりにくもったクラウドサービスlocaler•4/9/25, 2:00 PM

Hi. I encountered a consistent issue when editing a row in a D1 table via the dashboard. Attempting to edit the following string causes a 404 Bad Request:

data.txt51B

く

くもりにくもったクラウドサービスlocaler•4/9/25, 2:00 PM

I'm not exactly sure which character is triggering the error — it might be due to a special character or even an invisible whitespace — but the issue is 100% reproducible when trying to update a row containing this exact value.

ZZack We're already having to work around multiple other issues that we didn't expect ...

Tait•4/9/25, 5:24 PM

I'd love to hear more. What other issues did you unexpectedly encounter?

TTait I'd love to hear more. What other issues did you unexpectedly encounter?

ZackOP•4/9/25, 5:46 PM

Bulk import API causing the database to be unusable (which is documented to be fair), the D1 API having severe rate limits (worked around by using our own worker as a proxy for SQL commands), some issue internally with D1 causing huge latency after we replaced a bulk of the data within the database that the team fixed

ZackOP•4/9/25, 5:47 PM

We are using D1 in a slightly abnormal way, as a lookup table basically, so you probably won't experience the same problems if you're using it as a more traditional database

Tait•4/9/25, 8:57 PM

Thank you!

くくもりにくもったクラウドサービスlocaler I'm not exactly sure which character is triggering the error — it might be due t...

gruntlord6•4/9/25, 10:24 PM

does that end up in the url of the edit page?

Ggruntlord6 does that end up in the url of the edit page?

く

くもりにくもったクラウドサービスlocaler•4/9/25, 10:36 PM

Yes, the issue happens when click "Update" on a row that contains that string the value ends up in the URL, and that's when the 400 error oCcurs.
So it seems like the string being placed in the URL is causing the problem.

gruntlord6•4/9/25, 10:56 PM

its likely a character that can't be in a url, like the one used as a space character, the browser is interpreting it as a space etc and that causes the 404

gruntlord6•4/9/25, 10:57 PM

can't say which specific one is the issue but I'd say thats the issue IMO

jed•4/10/25, 1:15 AM

out of curiosity, does anyone know how D1 accomplishes autoincrement given its distributed nature?

Jjed out of curiosity, does anyone know how D1 accomplishes autoincrement given its d...

James•4/10/25, 1:19 AM

It doesn’t - d1 isn’t distributed today. Your db exists in a single region

jed•4/10/25, 1:22 AM

ah, so the

--location

--location

CLI param isn't so much the primary location as it is the only location.

jed•4/10/25, 1:23 AM

good to know, thank you @James!

James•4/10/25, 1:24 AM

Currently yep. The team has talked a lot about read replicas though, so I wouldn’t be surprised if we see something there soon

jed•4/10/25, 1:25 AM

awesome, i'll save my question for then, ha.

gruntlord6•4/10/25, 2:51 AM

really hoping for more d1 news given the outerbase stuff but not holding my breath

Ggruntlord6 really hoping for more d1 news given the outerbase stuff but not holding my brea...

Isaac McFadyen•4/10/25, 4:05 AM

James•4/10/25, 7:31 AM

That would be pretty great, fingers crossed!

o7•4/10/25, 12:44 PM

noway read replicas

ZackOP•4/10/25, 12:46 PM

it's happening

Samantas5855•4/10/25, 12:57 PM

Are there any issues with D1 currently?

Samantas5855•4/10/25, 12:57 PM

I'm pushing stuff but they don't show up on the API

Larry•4/10/25, 1:02 PM

It has happened!!! https://developers.cloudflare.com/changelog/2025-04-10-d1-read-replication-beta/

Cloudflare Docs

D1 Read Replication Public Beta

Larry•4/10/25, 1:03 PM

Next question, when does this come to DO-based SQLite storage?

LLarry Next question, when does this come to DO-based SQLite storage?

lambrospetrou•4/10/25, 1:04 PM

D1 is the customer zero of DOs

So, once we see it's good enough with D1 and iron out some things, DOs will get some version of it too.

LLarry It has happened!!! https://developers.cloudflare.com/changelog/2025-04-10-d1-rea...

lambrospetrou•4/10/25, 1:06 PM

I would encourage people to try this out, and please give us feedback. Tell us cases it broke for you, or didn't do what you expected. Note that it's "beta". We are aware of some specific issues and are fixed as we speak to be rolled out over the next few weeks, so do your own testing to see if fits your workloads too.

cosbgn•4/10/25, 1:08 PM

IMO at least for me the session API adds waaay too much work to bother implement it. Ideally I was hoping cf would just figure this out like turso does

cosbgn•4/10/25, 1:09 PM

Right now I simply use cache & KV which give me pretty much the same results but it's simpler and easier to understand (I know it's eventually consistent, but that works for me)

Ccosbgn Right now I simply use cache & KV which give me pretty much the same results but...

lambrospetrou•4/10/25, 1:11 PM

If you are fine with normal eventual cache, then keep with that, and it fits better. Global consistency has tradeoffs, either exposed to the user and you decide the queries you want consistent, or it's always done implicitly and then you get slower performance across the board.

Ccosbgn IMO at least for me the session API adds waaay too much work to bother implement...

lambrospetrou•4/10/25, 1:11 PM

Which Turso feature specifically? I am aware of the embedded replicas but that's a whole other thing/use-case.

cosbgn•4/10/25, 1:14 PM

They used to have multi region replicas, but from the homepage it looks like they discontinue it

Ccosbgn IMO at least for me the session API adds waaay too much work to bother implement...

Larry•4/10/25, 1:19 PM

In what way is it too much work? It looks like it's a few lines of code tweak? That may be more work than continuing to use your own cache+KV solution, but for anyone who hasn't already built that, it seems like adopting the session API is a lot LESS work than building something like you've done.

cosbgn•4/10/25, 1:29 PM

Maybe it's actually very little work, I've just quickly looked at the api, but it seems like I need to read incoming headers, tweak how drizzle interacts with the db, set a new header, etc.

With KV, I just do if (kv) return kn else return d1

The other advantage is that is not vendor lock in, I can switch to any other sqlite provider (like turso) with a single line change

cosbgn•4/10/25, 1:29 PM

but anyways, I wasn't really complaining, I love d1 and cloudflare, was more like my personal feedback

Given that will presumably replicate over time, and increase the PoP's it exists in, replacing the e

Similar Threads

Similar Threads

Similar Threads