ci: Add ibc-rs/Hermes integration test #35

thanethomson · 2021-12-22T22:43:52Z

See the README for details as to what this integration test does.

One major drawback to the generalized approach I've taken is that it's pretty slow to execute. In follow-up issues we can try to figure out ways to accelerate its execution (especially i.t.o. stashing build artifacts).

In terms of the large PR, half of it's just a Tendermint config.toml file that I bake into the image.

Signed-off-by: Thane Thomson <connect@thanethomson.com>

adizere · 2021-12-23T08:51:53Z

ci/tests/update-channel.sh

@@ -0,0 +1,11 @@
+#!/bin/bash
+set -euo pipefail


Should this be called update-client (instead of channel)?

Not sure why I named it that. I think I may have been half asleep 😅 I've updated the whole test now to be called create-connection.sh, and it instead executes a whole create connection operation as per the work in #20.

adizere · 2021-12-23T08:54:04Z

ci/README.md

+   ID `basecoin-0`).
+7. If no `CMD` arguments are provided for the container, it will automatically
+   execute the [`update-channel.sh`](./tests/update-channel.sh) script, which
+   creates and updates an IBC channel between `basecoin-0` and `ibc-0`. If `CMD`


I guess the plan is to eventually go all the way up to channels, but at the moment we stop at clients because connection handshake fails (cf. https://github.com/informalsystems/ibc-rs/issues/1710)?

This is a good point, but ATM we've been pinning ibc deps to specific branches (e.g. basecoin/phase-4-1 or basecoin/phase-5) in Cargo.toml. This was done to account for missing non-SDK chain support in ibc-rs. Now that proof verification and other things like custom proof-spec support have been implemented, we can finally start to depend on ibc-rs releases (once the fix for https://github.com/informalsystems/ibc-rs/issues/1710 is merged).
That said, we still haven't merged #27 yet, so I suggest we only run hermes create connection for now.

ci/Dockerfile

adizere · 2021-12-23T09:02:31Z

ci/tests/update-channel.sh

+HERMES_BIN=${HERMES_BIN:-${HOME}/build/ibc-rs/release/hermes}
+
+echo "Creating client..."
+"${HERMES_BIN}" tx raw create-client basecoin-0 ibc-0


Can we also create a client in the opposite direction, namely a client hosted on ibc-0 for chain basecoin-0?

Probably:

"${HERMES_BIN}" tx raw create-client ibc-0 basecoin-0 "${HERMES_BIN}" tx raw update-client ibc-0 07-tendermint-0

I've just added a commit now to create a connection in the opposite direction, but @hu55a1n1 mentioned here that it may intermittently fail. Does that still hold true @hu55a1n1?

Sorry, that comment is outdated. 😅 The connection creation shouldn't fail now. 🤞

Awesome, looks like the bidirectional connection integration test passes 😁 👍

adizere · 2021-12-23T09:06:15Z

ci/Dockerfile

@@ -0,0 +1,45 @@
+ARG TENDERMINT_VERSION=0.34.9
+ARG GAIA_VERSION=4.2.1


It may be better to use Gaiad v5 or v6 (mainnet hub-4 is at v6).

I can currently only get this working for Gaiad v5.0.5 (that's the latest image we have from cephalopodequipment in the v5.x series). v6.0.0 fails at the point where I try to add the generated user key to the basecoin-0 chain via Hermes:

Configuring Hermes... 2021-12-23T14:10:10.979891Z TRACE ThreadId(01) deserialized the encoded pub key into a ProtoAny of type '/cosmos.crypto.secp256k1.PubKey' Success: Added key 'testkey' (cosmos1va7pcr6gt6p8vvyexrmzwrm6k8uuc7rq2z96w0) on chain ibc-0 Generating user key... {"name":"user","type":"local","address":"cosmos177vu4gpafpzd9f39kwc64m3tevk628dr3xe6d0","pubkey":"{\"@type\":\"/cosmos.crypto.secp256k1.PubKey\",\"key\":\"A1QTDg0L9pZeHs0WHf73tTjyUBa8UI2Z931CHjUCSCVN\"}","mnemonic":"safe defense winter glide ladder belt clip inform they forest utility record feature cannon deliver stomach will mean destroy border matter sleep gospel involve"} Adding user key to basecoin-0 chain... Error: error encoding key: EOF while parsing a value at line 1 column 0

I'm assuming the serialization format changed between Gaiad v5 and v6?

Just tested this and it seems gaiad-v0.6.0 writes the output of the keys add command to stderr instead of stdout, so the user_seed.json file is empty leading to this problem.

adizere · 2021-12-23T12:10:35Z

I tried to run locally

docker run --rm -it \
     -v `pwd`:/src/basecoin-rs \
     -v /Users/adi/Hammers/ibc-rs/:/src/ibc-rs \
     informaldev/basecoin-rs-ci

and got this error:

Creating client...
2021-12-23T12:03:25.432795Z INFO ThreadId(01) using default configuration from '/basecoin/.hermes/config.toml'
2021-12-23T12:03:25.453362Z TRACE ThreadId(04) light client verification trusted=0-8 target=0-8
2021-12-23T12:03:25.461979Z DEBUG ThreadId(07) send_messages_and_wait_commit with 1 messages
2021-12-23T12:03:25.485075Z DEBUG ThreadId(07) [basecoin-0] send_tx: retrieved account sequence=0 number=0
2021-12-23T12:03:25.485991Z DEBUG ThreadId(07) [basecoin-0] send_tx: sending 1 messages using account sequence 0
2021-12-23T12:03:25.486916Z DEBUG ThreadId(07) [basecoin-0] send_tx: max fee, for use in tx simulation: Fee { amount: "401stake", gas_limit: 400000 }
2021-12-23T12:03:25.506339Z WARN ThreadId(07) [basecoin-0] estimate_gas: tx simulation successful but no gas amount used was returned, falling back on default gas: 400000
2021-12-23T12:03:25.506765Z DEBUG ThreadId(07) [basecoin-0] send_tx: using 400000 gas, fee Fee { amount: "401stake", gas_limit: 400000 }
2021-12-23T12:03:25.512568Z DEBUG ThreadId(07) [basecoin-0] send_tx: broadcast_tx_sync: Response { code: Ok, data: Data([]), log: Log(""), hash: transaction::Hash(225031418EE1D77A587E7C345BF92133D78752AD6A629CBF29370833F723A774) }
2021-12-23T12:03:25.512664Z INFO ThreadId(07) [basecoin-0] waiting for commit of tx hashes(s) 225031418EE1D77A587E7C345BF92133D78752AD6A629CBF29370833F723A774
2021-12-23T12:03:26.020330Z TRACE ThreadId(07) [basecoin-0] wait_for_block_commits: retrieved 1 tx results after 2 tries (306ms)
Success: ChainError(
"deliver_tx for 225031418EE1D77A587E7C345BF92133D78752AD6A629CBF29370833F723A774 reports error: code=Err(2), log=Log("deliver failed with error: no module could handle specified message\n\nLocation:\n /usr/local/cargo/registry/src/github.com-1ecc6299db9ec823/flex-error-0.4.4/src/tracer_impl/eyre.rs:10:9")",
)

Note that my ibc-rs directory was checked-out on master. I wonder what did I do wrong.

Signed-off-by: Thane Thomson <connect@thanethomson.com>

thanethomson · 2021-12-23T14:20:32Z

Note that my ibc-rs directory was checked-out on master. I wonder what did I do wrong.

Check out branch basecoin/phase-4-1 first and try again? This is the one I currently use in the integration test in CI until such time that all the basecoin-related work's been merged into ibc-rs.

thanethomson · 2021-12-23T14:23:22Z

ci/entrypoint.sh

+echo "Generating user key..."
+gaiad keys add user --keyring-backend="test" --output json > "${HOME}/user_seed.json"
+echo "Adding user key to basecoin-0 chain..."
+"${HERMES_BIN}" -c "${HERMES_CONFIG}" keys add basecoin-0 -f "${HOME}/user_seed.json"


This command specifically fails for Gaiad v6.0.0, and where I suspect the serialization in ibc-rs may need to be updated. cc @adizere @hu55a1n1

Signed-off-by: Thane Thomson <connect@thanethomson.com>

hu55a1n1

Thanks, @thanethomson! Awesome stuff! 💯👌

thanethomson added 3 commits December 22, 2021 17:43

Add first draft integration testing config

dfa67f0

Signed-off-by: Thane Thomson <connect@thanethomson.com>

Remove interactive flag from integration test

1f6c905

Signed-off-by: Thane Thomson <connect@thanethomson.com>

Fix checkout location for ibc-rs source

f74b3d0

Signed-off-by: Thane Thomson <connect@thanethomson.com>

thanethomson requested a review from hu55a1n1 December 22, 2021 23:05

thanethomson marked this pull request as ready for review December 22, 2021 23:05

thanethomson changed the title ~~ci: Add first draft integration testing config~~ ci: Add ibc-rs/Hermes integration test Dec 22, 2021

adizere approved these changes Dec 23, 2021

View reviewed changes

adizere reviewed Dec 23, 2021

View reviewed changes

thanethomson added 4 commits December 23, 2021 08:39

Replace test with create-connection (from phase 4)

8de5b64

Signed-off-by: Thane Thomson <connect@thanethomson.com>

Update to Gaia v5.0.5 and Tendermint v0.34.11

18f6bd2

Signed-off-by: Thane Thomson <connect@thanethomson.com>

Add more logging output for debugging purposes

e8635db

Signed-off-by: Thane Thomson <connect@thanethomson.com>

Update readme to reflect test script name change

abd1c01

Signed-off-by: Thane Thomson <connect@thanethomson.com>

thanethomson commented Dec 23, 2021

View reviewed changes

thanethomson mentioned this pull request Dec 23, 2021

ci: Bake Hermes binary into integration test #36

Closed

Create bidirectional connection in integration test

d8d8220

Signed-off-by: Thane Thomson <connect@thanethomson.com>

hu55a1n1 approved these changes Dec 24, 2021

View reviewed changes

hu55a1n1 merged commit 920a7c4 into main Jan 5, 2022

hu55a1n1 deleted the thane/17-integration-testing branch January 5, 2022 15:40

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

ci: Add ibc-rs/Hermes integration test #35

ci: Add ibc-rs/Hermes integration test #35

thanethomson commented Dec 22, 2021 •

edited

Loading

adizere Dec 23, 2021

thanethomson Dec 23, 2021

adizere Dec 23, 2021

hu55a1n1 Dec 23, 2021

adizere Dec 23, 2021

thanethomson Dec 23, 2021 •

edited

Loading

hu55a1n1 Dec 23, 2021

thanethomson Dec 23, 2021

adizere Dec 23, 2021

thanethomson Dec 23, 2021

hu55a1n1 Dec 24, 2021

adizere commented Dec 23, 2021

thanethomson commented Dec 23, 2021 •

edited

Loading

thanethomson Dec 23, 2021 •

edited

Loading

hu55a1n1 left a comment

		@@ -0,0 +1,45 @@
		ARG TENDERMINT_VERSION=0.34.9
		ARG GAIA_VERSION=4.2.1

ci: Add ibc-rs/Hermes integration test #35

ci: Add ibc-rs/Hermes integration test #35

Conversation

thanethomson commented Dec 22, 2021 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

thanethomson Dec 23, 2021 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

adizere commented Dec 23, 2021

thanethomson commented Dec 23, 2021 • edited Loading

thanethomson Dec 23, 2021 • edited Loading

Choose a reason for hiding this comment

hu55a1n1 left a comment

Choose a reason for hiding this comment

thanethomson commented Dec 22, 2021 •

edited

Loading

thanethomson Dec 23, 2021 •

edited

Loading

thanethomson commented Dec 23, 2021 •

edited

Loading

thanethomson Dec 23, 2021 •

edited

Loading