Ok, merci pour cette info que c’est un choix perso. Je vais avoir un autre choix perso ><
J’ai un peu cherché, voici le guide de bonnes pratiques pour Polkadot :
Par contre, je vois pas grand chose sur l’aspect réseau 
1 Like
Here is my ansible config for a Duniter node based on its role (mirror / archive / validator):
# Duniter {{role}} node only
version: "3.5"
services:
duniter-{{role}}:
image: {{image}}
restart: unless-stopped
ports:
# prometheus endpoint
- 127.0.0.1:9615:9615
# rpc via http
- 127.0.0.1:9933:9933
# rpc via websocket
- 127.0.0.1:9944:9944
# public p2p endpoint
- 30333:30333
volumes:
- duniter-data:/var/lib/duniter/
environment:
- DUNITER_CHAIN_NAME=gdev
- DUNITER_NODE_NAME={{duniter_node_name}}-{{role}}
{% if role=="validator" -%}
- DUNITER_VALIDATOR=true
{% endif -%}
- DUNITER_PRUNING_PROFILE={% if role=="mirror" %}default{% elif role=="archive" %}archive{% elif role=="validator" %}light{% endif +%}
- DUNITER_PUBLIC_ADDR=/dns/{{domain}}/tcp/30333
- DUNITER_LISTEN_ADDR=/ip4/0.0.0.0/tcp/30333
volumes:
duniter-data:
This way, I can update the docker image quite easily on multiple dockerfiles 
2 Likes
J’ai l’impression que mon nœud est bloqué dans sa synchro, ça fait 2j que mes essais mènent à ce blocage plus ou moins rapidement (ici je suis au bloc 1,194,752).
Cela vous est-il déjà arrivé ? Que faites-vous pour contourner ?
Édit : bon j’ai redémarré ça semble être débloqué.
Je n’ai eu de que très rares blocages. En fait un seul il me semble. Et c’est arrivé quand j’ai poussé mon serveur un peu trop loin en charge avec d’autres tests de services en parallèle. Sinon ça tourne tout seul avec uptime de plusieurs semaines sans souci.
@davidbp845 you might be interested in having a Duniter v2 node on the ĞDev5 network. The documentation is available here: Duniter | [alpha] Duniter v2. Tell me what you think about it!
Thanks a lot Hugo. I’ll read it and hope I can do it.
I’m experiencing trouble with my v1 node now, and first I would like to see if I can fix this. Then I will try to have a v2 one!
1 Like
Summary for @guenoel of the smith:
smith declared in genesis
| pseudo | identity index | address |
|---|---|---|
| poka | 6951 | 5CQ8T4qpbYJq7uVsxGPQ5q2df7x3Wa4aRY6HUWMBYjfLZhnn |
| elois | 6317 | 5CtmAELWQ6FGDrAfsfVbGe4CXW4YrpNRTVU27eWzSE6J8ocU |
| HugoTrentesaux | 2453 et 2457 | 5Dq8xjvkmbz7q4g2LbZgyExD26VSCutfEc6n4W4AfQeVHZqz |
| tuxmain | 7139 | 5D2DnScFpxoEUXDwZbJH18tRsQMygBSh1F6YCcWvTYzKY2W7 |
| 1000i100 | 1 | 5CCrBS67BrpBx3ihGHc72HZp3eHHbETxWFuNfwbbdoGSJFN8 |
| ManUtopiK | 3594 et 3595 | 5DUjwHRqPayt3tAZk1fqEgU99xZB9jzBHKy2sMSTNcc7m9D1 |
(pour les identités en double, cf Sanity tests - #6 by HugoTrentesaux)
smith membership requests
| bloc | pseudo | identity index | address |
|---|---|---|---|
| 632386 | pini | 7228 | 5GBVhdJUdsGhxozu6R8X6x2pTZvuuW46s7JSU4tiW7Zd3WmY |
| 917080 | hugo | 2457 | 5Dq8xjvkmbz7q4g2LbZgyExD26VSCutfEc6n4W4AfQeVHZqz |
| 917151 | poka | 6951 | 5CQ8T4qpbYJq7uVsxGPQ5q2df7x3Wa4aRY6HUWMBYjfLZhnn |
| 1491634 | AAAAA | 7237 | 5E58CLNXPFpBWrygp3ry8HyS2hiuQpRuXbZFUaTq7q9gG3Vv |
| 1606142 | vit | 7174 | 5FH48744BHgNoLBe8syGXbTEnSpGhp8ttdAKW4MWcR7TUKai |
2 Likes
We only have three smith on the ĞDev network (@daigongen, @tuxmain, @Pini) and the blockchain is experiencing latencies because @poka is offline. @vit and @HugoTrentesaux should become smith again, and we should train other smith.
Each time I restart the Validator node it stays stuck at an old block.
Dont know how to unlock this situation.
Is there an option to tell him which node to sync to or force re-sync like V1 ?
This is weird. I didn’t experience this at all with my nodes.
Try to remove its data and sync again. It will sync with the bootnodes in the genesis file.
Done.
Stuck to block 0 now…
home_duniter_validator.1.wziabozcgh94@z68-gen3 | 2023-06-07 13:27:44 ❌ Error while dialing /dns/telemetry.polkadot.io/tcp/443/x-parity-wss/%2Fsubmit%2F: Custom { kind: Other, error: Timeout }
home_duniter_validator.1.wziabozcgh94@z68-gen3 | 2023-06-07 13:27:49 💤 Idle (0 peers), best: #0 (0xf9bb…0dfc), finalized #0 (0xf9bb…0dfc), ⬇ 0 ⬆ 0
home_duniter_validator.1.wziabozcgh94@z68-gen3 | 2023-06-07 13:27:54 💤 Idle (0 peers), best: #0 (0xf9bb…0dfc), finalized #0 (0xf9bb…0dfc), ⬇ 0 ⬆ 0
Is there some ports I need to open/forward on my box ?
Ports 30334, 9616 are open.
ports:
# Prometheus endpoint
- 9616:9615
# rpc via http
# - 9933:9933
# rpc via websocket
- 127.0.0.1:9945:9944
# p2p
- 30334:30333
Do I need to do something for SSL on ngynx ?
My RPC done is doing well by the way…
[EDIT]
Added ;
- "DUNITER_PUBLIC_ADDR=/dns/{{domain_name}}/tcp/30334"
- DUNITER_LISTEN_ADDR=/ip4/0.0.0.0/tcp/30333
and restarted, with no luck…
I just saw it, my mini server didn’t start automatically after yesterday power outage, i will change that.
We need external monitoring with email alert or this kind of stuff.
I started from scratch a new validator node with image sha-2be3f534. A lot of endpoints are unavailable, I will publish a new image with correct bootnodes. My node is currently synchronizing: Polkadot Telemetry
Normal synchronisation is quite slow (about 50 minutes for 500,000 blocks), so I tried warp sync by adding to my docker-compose.yml:
command:
- "--sync=warp"
I got about 100,000 blocks in 5 minutes which is only twice as fast. Did anybody tried this synchronisation mode?
1 Like
I didn’t try. This stackexchange post says:
Warp sync - initially download the best finalized block and the whole state at this block, and finality proofs (currently from GRANDPA) for all validator set transitions from genesis up until the best finalized block. After the initial warp sync is done (and therefore the node is ready to import the latest blocks) historical blocks can be downloaded in the background.
I did my smith membership request using Ǧcli. It is on block 2,506,488.
# gcli command I used
$ gcli --network local --secret "xxx yyy zzz" smith request gdev.trentesaux.fr
transaction submitted to the network, waiting 6 seconds...
MembershipRequested(7238)
The option --network local tells Ğcli to connect to a local Duniter node, in this case a ssh bridge to my validator node with unsafe RPC unabled (for rotate keys).
# ssh bridge
ssh -NL 9944:localhost:9944 gdev.trentesaux.fr
As soon as we add the indexing of smith membership requests to the indexer, we will be able to track them in @guenoel’s monitoring tool gecko_web, and even to submit smith certs directly in this app using polkadotjs extension. (for the moment, you can see my old identity index 2457 which is hardcoded, but this is a bug of Ğdev, see ĞDev -- j'ai perdu mon identité and Sanity tests - #6 by HugoTrentesaux).
So please @smiths-GDev can you certify me? (identity number 7238)
The gcli command is:
# gcli command to certify
gcli --network gdev --secret "your secret phrase" smith cert 7238
# if I run this, I get `CannotCertifySelf`
But at the moment gcli is not released, so use polkadotjs app instead:
Thanks to @poka @tuxmain and @Pini I got my tree smith certs and got membership on block 2514380.
I was successfully able to go online with gcli:
$ gcli smith go-online
transaction submitted to the network, waiting 6 seconds...
MemberGoOnline(7238)
3 Likes
Salut à tous !
Bon, j’avais tenté une fois sans succès, et j’aimerais tenter une nouvelle fois d’installer un noeud.
Est-ce déjà ya un doc qq part pour comprendre le fonctionnement de duniter v2 (le fait qu’il y ait deux types de noeuds, par ex. et d’autres choses intéressantes à savoir pour faire tourner un noeud) ?
Je suis assez limité sur mon serveur infomaniak (j’ai pas réussi à installer un noeud miroir), donc je sais pas trop si je vais réussir…
Merci ! 
Ya un moyen d’installer docker “localement”, genre dans le home pour avoir les droits de le faire ?
Je n’ai pas vu que c’était possible sur internet, mais peut-être que des spécialistes auront une idée…
Merci !