Joystream Stats a3fe4e8e58 KPI 36.OP-1: content migration test		2 yıl önce
..
.env	df44c2e354 KPI 36.OP-1: query-node setup	2 yıl önce
Commands.md	e9472ecca2 KPI 36.OP-1: nginx setup	2 yıl önce
README.md	a3fe4e8e58 KPI 36.OP-1: content migration test	2 yıl önce
deploy-qn.sh	df44c2e354 KPI 36.OP-1: query-node setup	2 yıl önce
import-qn-data.sh	df44c2e354 KPI 36.OP-1: query-node setup	2 yıl önce
reset-qn.sh	df44c2e354 KPI 36.OP-1: query-node setup	2 yıl önce

KPI 36.OP-1

The goal is to install a GIZA validator, query node, storage provider and distributor on a fresh VPS.

Setup

Node

See Network deployment

Query-node

Edit: Some instructions have been simplified since the test (README olympia QN).

Download these to ~/joystream (or your repository root):

Adapt JOYSTREAM_NODE_WS in .env and import-qn-data.sh if necessary.

cd ~/joystream/query-node
yarn build
nano manifest.yml
# change second to last line to:
---
           height: "[13876,13876]"
---
cd ..
sudo docker-compose up
chmod +x import-qn-data.sh deploy-qn.sh reset-qn.sh
./reset-qn.sh

Storage

See storage-node.

Distributor

See distributor-node.

Pioneer

cd ~/joystream/pioneer
yarn
yarn build:www
rsync -aP packages/apps/build/ /var/www/pioneer

To pre-select your endpoint, edit packages/apps-config/src/settings/endpoints.js and add

  {
    info: 'Joystream Giza',
    text: t('giza-l1dev', 'Joystream Giza (l1dev)', { ns: 'apps-config' }),
    value: 'wss://giza-l1dev.joystream.app/rpc'
  },

as first array entry for createLive (line 23). Run yarn build:www again and you should see:

i18next-scanner: Added a new translation key { "giza-l1dev": "Joystream Giza (l1dev)" } to "packages/apps/public/locales/en/apps-config.json"

Alternatively it is possible to set WS_URL during build time: WS_URL=wss://myendpoint yarn build:www

Atlas

git clone htps://github.com/joystream/atlas
cd atlas
yarn
yarn build
rsync -aP dist/ /var/www/atlas

Set DNS record for atlas subdomain (depends on your registrar)

Create /etc/nginx/sites-available/atlas to configure nginx site

server {
listen 443 ssl;
server_name atlas.DOMAIN;

root /var/www/atlas;
index index.html;

include /etc/letsencrypt/options-ssl-nginx.conf;
ssl_dhparam /etc/letsencrypt/ssl-dhparams.pem;
}

Activate site

cd /etc/nginx/sites-enabled
ln -s ../sites-available/atlas

Request TLS certificate: certbot run, select the subdomain and hit enter.

Nginx

Install dependencies: apt install nginx certbot python3-certbot-nginx

Create /etc/nginx/sites-available/joysream to configure nginx site

server {
listen 80;
return 301 https://$host$request_uri;
}

server {
listen 443 ssl;
server_name giza-l1dev.joystream.app;

root /var/www/joystream; # build atlas or pioneer and move it there
index index.html;

location /rpc {
  proxy_pass http://localhost:9944;
  proxy_redirect off;
  proxy_http_version 1.1;
  proxy_set_header Upgrade $http_upgrade;
  proxy_set_header Connection "Upgrade";
  proxy_set_header Host $host;
  proxy_set_header    X-Real-IP $remote_addr;
}

# query node
location /graphql {
  rewrite /graphql(/.+) $1 break;
  proxy_pass http://localhost:8081;
  proxy_set_header Access-Control-Allow-Methods "GET, PUT, HEAD, OPTIONS";
}
location /@apollographql { rewrite (.*) /graphql$1 last; }

# storage provider
location /storage {
  rewrite /storage/?(.*) /$1 break;
  proxy_pass http://localhost:3333;
  proxy_set_header Access-Control-Allow-Methods "GET, PUT, HEAD, OPTIONS";
  proxy_set_header Access-Control-Allow-Headers "GET, PUT, HEAD, OPTIONS";
  client_max_body_size 50G;
}

# distributor
location /distributor {
  rewrite /distributor/?(.*) /$1 break;
  proxy_pass http://localhost:3334;
  proxy_set_header Access-Control-Allow-Methods "*";
  proxy_set_header Access-Control-Allow-Headers "*";
}

include /etc/letsencrypt/options-ssl-nginx.conf;
ssl_dhparam /etc/letsencrypt/ssl-dhparams.pem;
}

Activate site

cd /etc/nginx/sites-enabled
ln -s ../sites-available/joystream
rm default

Request TLS certificate: certbot run. Enter email address, accept the TOS to create a certbot account. The rest happens automatically. No need to add a redirect because it is already configured above.

Test:

open Pioneer and choose as custom endpoint: wss://giza-l1dev.joystream.app/rpc
https://giza-l1dev.joystream.app (if pioneer or atlas are set up)
https://giza-l1dev.joystream.app/graphql
https://giza-l1dev.joystream.app/storage/api/v1/version
https://giza-l1dev.joystream.app/storage/api/v1/state/data-objects
https://giza-l1dev.joystream.app/distributor/

Configuration

Apply as storage provider and distributor, save files to ~/keys/
Have the leads fill the openings
Get invited to a bucket

Verify group memberships

$ ~/joystream/cli/bin/run working-groups:overview

Current Group: storageProviders

___________________ Group lead ___________________

Member id:                    2010                                             
Member handle:                storage_lead_giza                                
Role account:                 5GHPwGqcVjt3KKrEUhkgLJaizRymegx5astt4fekGVmL7F4q 

____________________ Members _____________________

Worker id     Member id     Member handle         Stake            Earned     Role account                  
3             515           l1dev                 10.0000 JOY      0          5DNe6u...ATpk38               
2             2012          storage_2             100.0000 JOY     0          5DGQLX...Fai2MW               
1             2012          storage_2             100.0000 JOY     0          5H8CPz...hoWPtP               
0             2010          storage_lead_giza     100.0000 JOY     0          5GHPwG...mL7F4q

$ ~/joystream/cli/bin/run working-groups:overview -g distributors
Current Group: distributors

___________________ Group lead ___________________

Member id:                    2015                                             
Member handle:                dist_lead                                        
Role account:                 5GJWFApQiMqFjhiLWhxtVAjva7fVy7kYAf8mBgME4rGXX4Vy 

____________________ Members _____________________

Worker id     Member id     Member handle     Stake            Earned     Role account                  
2             515           l1dev             100.0000 JOY     0          5DCCUr...qjEvK9               
1             2016          dist_1            100.0000 JOY     0          5GQZGW...m9z7VA               
0             2015          dist_lead         100.0000 JOY     0          5GJWFA...GXX4Vy

Storage Provider

Accept invitation

$ ~/joystream/storage-node-v2/bin/run operator:accept-invitation -i 2 -w 3 -k ~/keys/giza-storage-l1dev.json -t 5DNe6ubsQhmcSZRZdBLJroe9qKThW9CsSGRKHiZx8pATpk38
2022-01-04 15:43:26:4326 info: Initialized runtime connection: ws://localhost:9944
2022-01-04 15:43:30:4330 info: Waiting for chain to be synced before proceeding.
2022-01-04 15:43:30:4330 info: Accepting pending storage bucket invitation...
2022-01-04 15:43:30:4330 debug: Sending storage.acceptStorageBucketInvitation extrinsic...
2022-01-04 15:43:36:4336 debug: Extrinsic successful!

set metadata: storage-node operator:set-metadata (example json - fields)

$ ~/joystream/storage-node-v2/bin/run operator:set-metadata -i 2 -w 3 -j ~/bucket-2-metadata.json -k ~/keys/giza-storage-l1dev.json
2022-01-04 15:46:21:4621 info: Initialized runtime connection: ws://localhost:9944
2022-01-04 15:46:24:4624 info: Waiting for chain to be synced before proceeding.
2022-01-04 15:46:24:4624 info: Setting the storage operator metadata...
2022-01-04 15:46:24:4624 debug: Sending storage.setStorageOperatorMetadata extrinsic...
2022-01-04 15:46:30:4630 debug: Extrinsic successful!

Distributor

See docs/node

Accept invitaiton

~/joystream/distributor-node/bin/run operator:accept-invitation -B 2:0 -w 2 -c ~/joystream/distributor-node/config.yml
2022-01-04 16:08:16:816 CLI info: Accepting distribution bucket operator invitation...
{
"bucketId": {
    "distribution_bucket_family_id": "2",
    "distribution_bucket_index": "0"
},
"workerId": 2
}
? Tx fee of 0 will be deduced from you account, do you confirm the transfer? (y/N) y
? Tx fee of 0 will be deduced from you account, do you confirm the transfer? (y/N) y? Tx fee of 0 will be deduced from you account, do you confirm the transfer? Yes
2022-01-04 16:08:22:822 SubstrateApi info: Sending storage.acceptDistributionBucketInvitation extrinsic from 5DCCUrwUZuqBHMMcaswK5rk1DpFNGvuZvoZHWiHYNyqjEvK9
2022-01-04 16:08:24:824 CLI info: Invitation succesfully accepted!

set metadata (or use the API)

~/joystream/distributor-node/bin/run operator:set-metadata -B 2:0 -w 2 -c ~/joystream/distributor-node/config.yml -i ~/distributor-metadata.json
2022-01-04 16:25:29:2529 CLI info: Setting bucket operator metadata...
{
"bucketId": {
    "distribution_bucket_family_id": "2",
    "distribution_bucket_index": "0"
},
"workerId": 2,
"metadata": {
    "endpoint": "https://giza-l1dev.joystream.app/distributor",
    "location": {
        "countryCode": "US-NJ",
        "city": "Newark",
        "coordinates": {
            "latitude": 40,
            "longitude": 74
        }
    },
    "extra": "Welcome to Joystream GIZA - America branch! -- operated by l1dev (dev@joystreamstats.live)"
}
}
? Tx fee of 0 will be deduced from you account, do you confirm the transfer? (y/N) y
? Tx fee of 0 will be deduced from you account, do you confirm the transfer? (y/N) y? Tx fee of 0 will be deduced from you account, do you confirm the transfer? Yes
2022-01-04 16:25:39:2539 SubstrateApi info: Sending storage.setDistributionOperatorMetadata extrinsic from 5DCCUrwUZuqBHMMcaswK5rk1DpFNGvuZvoZHWiHYNyqjEvK9
2022-01-04 16:25:42:2542 CLI info: Bucket operator metadata succesfully set/updated!

Runtime

Staking and Rewards (WGs)

For each group (incl. Leads), check that:

staking and unstaking works as expected

rewards are paid out as they should

Validator Rewards

The runtime upgrade included a change in these params. For each era, get:

total issuance

total stake

total reward

~blockheight

Block	Era	Issuance	Stake	% Staked	Era Reward
0	0	574460220	15000	0.0	3275
6000	10	574460120	15000	0.0	3275
6600	11	574455020	15000	0.0	5698
7200	12	1000000000	200000000	20.0	58841
13200	22	1000000000	200000000	20.0	58841
13800	23	999999800	200000000	20.0	14805
14400	24	999999700	200000000	20.0	14780
15000	25	999999700	100007500	10.0	10250
15600	26	999999600	100007500	10.0	10250
16200	27	999999600	100007500	10.0	10250
16800	28	999999600	200000000	20.0	14805
151800	253	999999500	200000000	20.0	14805
152400	254	999999500	200000000	20.0	14780
153000	255	999999500	200000000	20.0	14805
161400	269	999999400	200000000	20.0	14805
162000	270	999999400	200500000	20.1	14828
187800	313	999999400	200500000	20.1	14828
188400	314	999999400	200500000	20.1	15397
192000	320	1099991546	200500000	18.2	15397

Storage

The storage tests are delayed to improve the content migration script (for non-sudo uploads and to support dynamic channel-bucket-provider assignment - see (PR #3034).

What happens when a storage node runs out of (local) storage?
- eg. the bucket is configured to accept 100GB, whereas the node runs out at 80GB
What happens when a bucket is full (ref. bucket config on-chain) and you upload more from a channel where it's assigned to accept from?
What happens when a bucket is full (ref. bucket config on-chain) and you assign a new bag to it?
What happens when a SP node, that is the only source of some data goes down?
- Cases:
- No distributor has it (or is assigned): Try to assign the channel bag to another SP, and use the re-upload command (cli and/or atlas)
- One or more distributors has it (and is assigned): Try to assign the channel bag to another SP, and use the re-upload command (cli and/or atlas)
- One or more distributors has it (and is assigned): Try to assign the channel bag to another SP, and try to fetch 5it from the distributor
How does the blacklist work?
- Cases:
- One distributor has (and is assigned) the file. asset it get censored in atlas?
  - What happens if you upload the asset again?
- No distributor has the file (or is assigned). When one gets assigned it, does it get it?
  - What happens if you upload the asset again?

CLI

Check that doing --help for all commands

provides a useful description

everything is correct

syntax is consistent

See Commands

Distributor

What happens when a distributor node runs out of storage?
- eg. the bucket is configured to accept 100GB, whereas the node runs out at 80GB

CLI

Check that doing --help for all commands

provides a useful description

everything is correct

syntax is consistent

all the examples are correct

See Commands

No help:

start
dev:batchUpload
node:*
leader:cancel-invitation
leader:create-bucket-family
leader:delete-bucket-family
leader:invite-bucket-operator
leader:remove-bucket-operator
leader:set-bucket-family-metadata
leader:set-buckets-per-bag-limit
leader:update-bucket-mode

Hours

Date	Hours	Task
Jan 2 13-17	4	start node, storage provider, distributor
Jan 4 15-18	3	storage provider and distributor configuration
Jan 4 19-22	3	check commands, validator eras
Jan 5 20-23	3	nginx config
Jan 6 18-23	5	content creation
Jan 7 12-15	3	query-node deployment
Jan 10 6-11	5	research joystream issue updates
Jan 13 6-12	6	content-migration script, PR #3034
Jan 14 2-7	5	content-migration script, PR #3034
	37

README.md

KPI 36.OP-1

Setup

Node

Query-node

Storage

Distributor

Pioneer

Atlas

Nginx

Configuration

Storage Provider

Distributor

Runtime

Staking and Rewards (WGs)

Validator Rewards

Storage

CLI

Distributor

CLI

Hours