Deploy `logos.test` fleet META issue

[kaiserd]

Deploy logos-nodes for the logos.test fleet.

The logos node should use the following configs for logos.test. For now, these are the same as described in the logos.dev setup:

• messaging module: Deploy logos.dev fleet - Delivery Components #1
• storage module: Deploy logos.dev fleet - Storage Components #8
• blockchain module: Deploy logos.dev fleet - Blockchain component #10

The difference between logos.test and logos.dev: in logos.test each logos node runs all these modules; in logos.dev each logos node only runs one of these modules.

Additional components for modules are covered in separate issues:

• Logos Blockchain legacy nodes
• Postgres DB node for Logos Delivery (see Deploy logos.dev fleet - Delivery Components #1 )

In addition to the logos.test fleet, which will be part of a public testnet, we would like a separate internal .dev fleet.

Module specific config for logos.dev is listed in the following issues:

• Deploy logos.dev fleet - Delivery Components #1
• Deploy logos.dev fleet - Storage Components #8
• Deploy logos.dev fleet - Blockchain component #10

Domain name update: #17

[igor-sirotin]

Agreed fleet structure

• logos.test

  The official Testnet fleet. Used by external people, starting from PSF.

  Consists of nodes running Logos Node with all components included:

  • node-01...logos.test
  • node-02...logos.test
  • ...

  Can be scaled with Logos Node running individual components on demand.

[davidrusu]

@jakubgs for the testnet fleet, can we have the same deployment process as in the existing devnet fleet? i.e. we push to a branch, that triggers the redeployment.

Also, currently we are using the testnet branch for the devnet fleet. Can we switch that around?

devnet branch -> existing devnet fleet
testnet branch -> new logos.test fleet

[igor-sirotin]

For Messaging, can we please have the same setup as we currently have in logos.dev

[kaiserd]

@jakubgs for the testnet fleet, can we have the same deployment process as in the existing devnet fleet? i.e. we push to a branch, that triggers the redeployment.

Also, currently we are using the testnet branch for the devnet fleet. Can we switch that around?

devnet branch -> existing devnet fleet testnet branch -> new logos.test fleet

Just to clarify, you are referring the the current blockchain internal devnet, not the logos.dev fleet?
I'd suggest we also deploy logos nodes running the blockchain module in the logos.dev

---

You comment covers the running legacy blockchain nodes in the logos fleets.
Since we would mainly run logos nodes: What would be the config for the blockchain module running in the logos nodes deployed on logos.test and logos.dev?`

[davidrusu]

Ah I see, ok, disregard previous comment!

@jakubgs we've put up this PR on logos-docker that shows the expected configuration setup:
logos-co/logos-docker#6

But in a nutshell

1. First generate the config with latest initial peers:

LOGOS_BLOCKCHAIN_PARAMETERS='{\
  "initial_peers": [\
    "/ip4/65.109.51.37/udp/3001/quic-v1/p2p/12D3KooWMwekH9M34FQXX6jPX4XYyNDgaCw68g4y7hqW2bv4UU9h",\
    "/ip4/65.109.51.37/udp/3002/quic-v1/p2p/12D3KooWEyJNQzmnfWcaRzCBv1ebqXMAWJecnBWmzp2AF6gk28PQ",\
    "/ip4/65.109.51.37/udp/3003/quic-v1/p2p/12D3KooWMrcbQQwT2dpNDiuLUdGESoqR3u2sJ8gAHHo9JPQH7yHb",\
    "/ip4/65.109.51.37/udp/3000/quic-v1/p2p/12D3KooWNMixUYkNuvXrNwnsmK2ZFRKNm48kF2qqtd9Du7Ls1yQe"\
  ],\
  "output": "/etc/logos/blockchain/node_config.yaml"\
}'

logoscore \
  -m ./modules \
  --load-modules liblogos_blockchain_module \
  -c "liblogos_blockchain_module.generate_user_config_from_str('${LOGOS_BLOCKCHAIN_PARAMETERS}')"

this will write a config to /etc/logos/blockchain/node_config.yaml

2. start the node

LOGOS_BLOCKCHAIN_DEPLOYMENT=devnet
LOGOS_BLOCKCHAIN_CONFIG_PATH=/etc/logos/blockchain/node_config.yaml # same as from the config gen step

logoscore \
  -m ./modules \
  --load-modules liblogos_blockchain_module \
  -c "liblogos_blockchain_module.start('${LOGOS_BLOCKCHAIN_CONFIG_PATH}', '${LOGOS_BLOCKCHAIN_DEPLOYMENT}')"

[jakubgs]

I don't believe we use logos-docker repo since it's just weirdly detached from what it actually referes to.

I believe we use compose from logos-liblogos(which will be renamed to logos-core), isn't that right @mendelskiv93 ?

[davidrusu]

Well, anyhow, the config is the same, independent of which docker image you use.

[mendelskiv93]

I don't believe we use logos-docker repo since it's just weirdly detached from what it actually referes to.

I believe we use compose from logos-liblogos(which will be renamed to logos-core), isn't that right @mendelskiv93 ?

Correct.

[jakubgs]

Also, why are we mixing up dev and test fleets in this issue?`

Please lets not confuse things further. If this meta issue is for logos.dev then we need to change the title.

[jakubgs]

Also, this issue is not for "blockchain module" as the description states, we need a separate issue for that.

[kaiserd]

For now logos.test and logos.dev would have the same module configs.

The difference between logos.test and logos.dev: in logos.test each logos node runs all these modules; in logos.dev each logos node only runs one of these modules.

Also, this issue is not for "blockchain module" as the description states, we need a separate issue for that

Yes. I'll link the issue as it is opened.

[kaiserd]

@iurimatias will add info on how to start the Logos node, using the configs linked in this issue.