Cannot upgrade my dApp - 503 Service Unavailable, unable to buffer body

Developers
1

As of 8/21/2025 I can no longer upgrade my dApp on the ICP mainnet,

canister id t5cvz-wiaaa-aaaaa-qag4q-cai

subnet snjp4-xlbw4-mnbog-ddwy6-6ckfd-2w5a2-eipqo-7l436-pxqkh-l6fuv-vae

WASM binary file is quiet weighty ~11mb but this -never- was an issue before, today after subnet was upgraded MOST of my dfx canister install calls FAIL with 503 error:

Error: Failed while trying to deploy canisters.
Caused by: Failed while trying to install all canisters.
Caused by: Failed to install wasm module to canister 'clown_mining_backend'.
Caused by: Failed during wasm installation call
Caused by: The replica returned an HTTP Error: Http Error: status 503 Service Unavailable, content type "text/plain; charset=utf-8", content: Unable to buffer body

I could try reducing WASM size, but this is very painful.

Any idea what happened? Why cannot I deploy my dApp?

On top of this dashboard says “undefined Canister”…

Issue is new and been plaguing me whole day, most of upgrades 9/10 gonna fail with this 503 error

2 Likes
2

Did you turn off your canister and turn it on again?

2 Likes
3

Is it 11 megs zipped? Wut

2 Likes
4

No. It’s the raw size of the wasm binary, looks like I wasn’t using gzip with dfx

But still… it never was an issue before, I think supporting bigger WASM files is important

1 Like
5

I just tried gzipping and got binary size down to 4.4MB, it still fails with Caused by: The replica returned an HTTP Error: Http Error: status 503 Service Unavailable, content type “text/plain; charset=utf-8”, content: Unable to buffer body from time to time, but success rate is higher

Something got broken on the node side I think, all this mess started right after today’s subnet upgrade

1 Like
6

Did the problem persist or did you manage to install all your canisters?

1 Like
7

Hi. I’ve just tried to deploy again several times - issue seems to be resolved

It works both with 4.4 gzipped WASM and with uncompressed WASM(~11mb) like it used to do.

I wonder was it intermittent issue or some kind of fix was deployed?

Any idea what was causing this?

1 Like
8

It must be related to this

1 Like
9

I’m seeing this error again and a lot, please what can I do except reducing the WASM size?

10

If compressing the Wasm artifacts helps, then that’s your best option.

You may also consider moving your dapp to another subnet, if this one is always busy.

11

Can this be helped on the node level? I think it’s important to support bigger WASM

Why are we being so restrictive with this?

12

Moving subnets is hard to do, I’d lose all the user data. Backing up on/off chain is hard(dApp has a lot of very complicated internal state)

I wish snapshots could be used to move canister state to another subnet/canister id, is this planned?

Also what happens when Funnai or somebody else joins and clogs new subnet?

1 Like
13

This is in the works right now. Up/downloading snapshots is close to being deployed, but I don’t know if that already allows uploading snapshots to different canister ids. But if not then that is the next thing to do for the ones working on it.

1 Like
14

Can this be helped on the node level? I think it’s important to support bigger WASM

Are you aware of the chunk upload feature? Are you using dfx other tooling?

And yes, canister migration (including ID) is coming soon.

1 Like
15

I’m using dfx 0.26.0

16

Wow all I can say is you’re an amazing dev man :high_voltage::heart::heart:

17