[Comms] Görli Testnet Contributors Call #4 "Istanbul"

[soc1c]

Görli Testnet Contributors Call #4

Date / Time

• Tuesday, September 17th, 2019
• 12pm UTC

Agenda

• Istanbul Hardfork Activation
  • Date/Time
  • Block number

All contributors and volunteers invited. Join Gitter for invite: gitter.im/goerli/testnet

cc @Souptacular

---

scroll down for meeting minutes: #60 (comment)

[soc1c]

I will propose block number 1561651 (12:00 pm UTC | Wednesday, October 30, 2019) for Görli Testnet. That is 6 weeks from now and well after Devcon 5.

[timbeiko]

IIRC, the goal was to have the upgrade on Oct 2nd, prior to devcon, for Goerli. (see previous decision)

If we do still want to fork Goerli at the same time as Ropsten, then I would propose 1400041, which should be around October 2nd.

[soc1c]

That's in two weeks. It will be a mess.

[roninkaizen]

i agree with soc1c; could we please discuss this tommorow during the call?
i am really willing and also interested in participating in testing
or pre-testing in close correspondance with Thomasz from nethermind.

[tkstanczak]

is the call happening?

[soc1c]

yes in 5 min https://us04web.zoom.us/j/388590582

[soc1c]

Protocol

• Afri welcomes everyone
  • Attendees: Adrian, Afri, Danno, Hudson, Tim, Tomasz, Vaibahv, Ronin
• Hudson starts with background on Istanbul planning and the testnet hardfork date Oct/2, 2019
  • Istanbul was estimated to hit mainnet in Oct
  • It was decided to have longer testnet testing phase though
  • It was decided to have testnet fork Oct/2
  • Görli has most clients, Ropsten has most traffic
  • No block number for any of the testnets proposed yet
  • all-core-devs assumed 2 weeks is enough for announcement
• Tomasz says Görli should not be the first testnet to fork
  • but in case everything goes well, it is also the best network to fork
  • actual state tests will come up after the actual hardfork
• Hudson previously suggested palindrome numbers for Ropsten and Görli
  • However, safety should be number 1 priority
  • Also there is Devcon 5 in Osaka coming up
• Afri explains his Oct/30 logic (Devcon)
  • two weeks is not enough to publish releases for all clients and ping all validators to upgrade
  • four weeks would be Devcon 5
  • six weeks is optimal as everyone is back from the conferences
• Hudson agrees it would be complicated before/during Devcon
  • But also sees some urgency and suggests to just let it run during Devcon
• Hudsom proposes Ropsten fork on Oct/2 and Görli on Oct/30
• Tomasz proposes bootstrapping a new Istanbul-only testnet
  • people were asking about it
  • Hudson approves this
  • Afri approves new testnet and Hudson's HF plans
  • Tomasz and Hudson agree with Oct/2 Ropsten fork and Oct/30 Görli fork
• Adrian asks whether clients are ready
  • Hudson says all major clients except for Aleth and Trinity merged Istanbul
• Danno (inaudible), discussing releases for Pantheon
• Hudson suggests to use 1561651 for Görli as Afri suggested
  • general agreement
• Tomasz: all clients needs releases
  • we need to start talking to all validators
• Afri concludes
• Bonus: Afri computes block time of 17.13 seconds on Ropsten and proposes 6485846 as fork block number for Ropsten Istanbul activation
  • goal should be doing one release for both testnets
  • Hudson will forward it to all-core-devs
• Afri again encourages to have one release for both testnets
• Danno discusses future course of action for hardforks
  • coordination between all-core-devs and Görli community
• Afri suggests a heads-up of at least 4 weeks for Görli hardforks
• Adrian kindly requests all validators to swap out the pantheon node for the new one:

  clique.propose("0x5dfe3772ceb73531357c3c747b6999cf59a59533", true)
  clique.propose("0x22ea9f6b28db76a7162054c05ed812deb2f519cd", false)
  

• Bye!

TL;DR

• Görli decision on 1561651 (Oct 30)
• Ropsten proposal 6485846 (Oct 2) for all-core-devs