Cloudflare Developers•9mo ago

how frequent?

CChaika how frequent?

Zack•4/10/25, 9:32 PM

Every request when I was testing

ChaikaOP•4/10/25, 9:33 PM

do you have replication enabled on the db? I see some on my replicated db, but none w/o

Zack•4/10/25, 9:33 PM

I can test again shortly

CChaika do you have replication enabled on the db? I see some on my replicated db, but n...

Zack•4/10/25, 9:33 PM

Yes, replicated, non replicated databases are totally error free

Zack•4/10/25, 9:33 PM

FWIW this isn't a big issue for me, but if it's more widespread I am happy to help with providing errors and retesting

Zack•4/10/25, 9:33 PM

I am just testing the replicated d1 in dev right now but it just won't work for any requests

Samantas5855•4/10/25, 9:54 PM

apparently my issues seem to be due to caching in pages functions

ZZack I am just testing the replicated d1 in dev right now but it just won't work for ...

Milan•4/10/25, 10:36 PM

Dumb question but wdym "in dev"?

MMilan Dumb question but wdym "in dev"?

Zack•4/10/25, 10:37 PM

As in, our dev environment

Zack•4/10/25, 10:37 PM

I've not pushed using replication into our production environment yet, so it's not a big issue for me

Zack•4/10/25, 10:37 PM

I am still experiencing the problems though

Milan•4/10/25, 10:37 PM

Like, running wrangler --dev remotewrangler --dev remote, or another namespace separate from your production namespace, but still deployed to cloudflares network?

Zack•4/10/25, 10:38 PM

All deployed to cloudflare, we deploy via terraform

Zack•4/10/25, 10:38 PM

Seperate database to production

Zack•4/10/25, 10:38 PM

It's "production" as far as you guys know

Milan•4/10/25, 10:41 PM

Thanks for clarifying!

MMilan Thanks for clarifying!

Zack•4/10/25, 10:46 PM

Some more info, I disable read replication on the database, redeployed the worker, everything worked fine. Enabled replication again, redeployed the worker, constant errors again

Zack•4/10/25, 10:46 PM

D1 ID: c60521e7-59bd-4186-81e7-9a676edf9b75
Incase you wanted to look into it. If you need the worker ID, just lmk what specifically you need

Aajgeiss0702 db id is `221100ba-5ab9-4cda-8cfa-69c17ea546ab` if it helps (account id is `f55b...

ajgeiss0702•4/10/25, 10:59 PM

i am also getting the issue occassionally

MMilan Thanks for clarifying!

Zack•4/10/25, 11:04 PM

Sorry to reply one more time, but I just wanted to say that I hadn't moved to use withSessionwithSession yet, might just be coincidence, but I just deployed the worker with that change, and now the replicated D1 database is working just fine

ZZack Sorry to reply one more time, but I just wanted to say that I hadn't moved to us...

Milan•4/10/25, 11:22 PM

I'm not familiar enough with D1's API to know whether this would cause problems or not, but can get back to you tomorrow

Zack•4/10/25, 11:23 PM

My understanding is that without using the session API it will only ever route to the primary instance of D1

Zack•4/10/25, 11:23 PM

But it shouldn't just fall over and constantly fail

Milan•4/10/25, 11:26 PM

I think that's probably right, it's possible the redeploy moved your primary somewhere else and the previous host was just having some weird ephemeral issue, but we'll figure it out and let you know

Zack•4/10/25, 11:27 PM

Sounds good

ZZack My understanding is that without using the session API it will only ever route t...

lambrospetrou•4/10/25, 11:32 PM

Without using the Sessions API (withSession()withSession()) all requests are sent to the primary explicitly. Under the hood the D1 binding has already been using the Sessions API for non replicated databases for months now, and it sends first-primaryfirst-primary when you don't use withSession()withSession() yourself.

Zack•4/10/25, 11:33 PM

Makes sense, to be more specific then, I now use withSession('first-unconstrained')withSession('first-unconstrained') which is working as I would expect, and not experiencing those same errors

ajgeiss0702•4/10/25, 11:56 PM

I switched to using the sessions api for one of my D1 DBs (that has replication enabled) and now almost every request is failing with the

D1_ERROR: Network connection lost.

D1_ERROR: Network connection lost.

error. Sometimes after a few seconds, sometimes within a second

ajgeiss0702•4/11/25, 12:00 AM

nvm now its fine again, seems very random when it happens

Aajgeiss0702 I switched to using the sessions api for one of my D1 DBs (that has replication ...

lambrospetrou•4/11/25, 12:16 AM

Network connection lost.Network connection lost. are transient errors reaching the location/datacenter of your database.

My suggestion is to add retries for some common errors. See an example in the D1 read replication demo of how to do retries: https://github.com/cloudflare/templates/blob/main/d1-starter-sessions-api-template/src/index.ts#L108
And how it's used: https://github.com/cloudflare/templates/blob/c624b31b63568adebcba1b63f600dbc81cf3a5fa/d1-starter-sessions-api-template/src/index.ts#L59

Llambrospetrou `Network connection lost.` are transient errors reaching the location/datacenter...

ajgeiss0702•4/11/25, 12:18 AM

I've gotten this error before and did retries there, but after enabling replication the error is happening way more than it did before (even though the dashboard shows that no replicas have been made yet)

ajgeiss0702•4/11/25, 12:18 AM

before, this error would only happen on some of my hottest paths on the most popular day of the week, but now its even happening today (which is a low day) on paths that don't get a ton of traffic

Aajgeiss0702 I've gotten this error before and did retries there, but after enabling replicat...

Zack•4/11/25, 12:18 AM

Have you moved your worker to use the sessions API?

ZZack Have you moved your worker to use the sessions API?

ajgeiss0702•4/11/25, 12:19 AM

at first I didnt, but now I have and it still happens occassionally.
When I first deployed the sessions api for one of my DBs, all requests failed with this error for a few minutes

Zack•4/11/25, 12:19 AM

What bookmark constraint are you using?

Aajgeiss0702 before, this error would only happen on some of my hottest paths on the most pop...

lambrospetrou•4/11/25, 12:20 AM

What's a "low day"? These errors are not dependent on your workload, they depend on the health of the datacenter your worker runs and of the database.

ZZack Have you moved your worker to use the sessions API?

lambrospetrou•4/11/25, 12:20 AM

That's irrelevant

ZZack What bookmark constraint are you using?

ajgeiss0702•4/11/25, 12:20 AM

im not specifying (which should default to first-unconstrainedfirst-unconstrained) just doing withSession()withSession()

Llambrospetrou That's irrelevant

Zack•4/11/25, 12:22 AM

Surely not if the query is always using 'firstPrimary' without reusing the session?

Zack•4/11/25, 12:22 AM

As then it will only ever hit the primary instance

Zack•4/11/25, 12:22 AM

Still shouldn't be seeing those errors presumably, but it'd explain why they only have a single instance count

ZZack Surely not if the query is always using 'firstPrimary' without reusing the sessi...

lambrospetrou•4/11/25, 12:23 AM

The discussion is about the network connection lost errors, not about what database instance is hit

Llambrospetrou What's a "low day"? These errors are not dependent on your workload, they depend...

ajgeiss0702•4/11/25, 12:24 AM

the site im testing this with is a site centered around an event that happens every friday evening, so any day other than friday has very little traffic compared to friday.
The "most popular" page that ive seen with this error has only had 45 requests in the last 24h, of which 6 have failed with this error.
there are other paths that have thrown this error, but thats the one with the most requests in the past 24h

ajgeiss0702•4/11/25, 12:27 AM

before enabling replicas this morning, there were 0 errors from any of these less popular pages. Now theres a few dozen

Zack•4/11/25, 12:40 AM

Alright, so after the success of using withSession('first-unconstrained')withSession('first-unconstrained') in dev, I decided to try and push that to prod. Now the database is entirely unresponsive and hasn't managed to serve a successful request for the last 10 minutes. Continually getting network connection lost errors

ZZack Alright, so after the success of using `withSession('first-unconstrained')` in d...

ajgeiss0702•4/11/25, 12:41 AM

that happened to me for one of my DBs too. It started working after a few minutes

Zack•4/11/25, 12:42 AM

I'm giving it a bit of time to recover from the sudden load, but can't really give it much longer, it's been ~15 minutes now

ajgeiss0702•4/11/25, 12:43 AM

looks like mine took about 17 minutes to recover

Zack•4/11/25, 12:43 AM

I'll give it another 5, then I will have to revert as we are serving entirely from cache atm, any misses are throwing 500's

Zack•4/11/25, 12:50 AM

Alright, rolling back. Database never recovered

how frequent?

Similar Threads

Similar Threads

Similar Threads