Try   HackMD

Merge Team Call - … - 2020-07-06

Where: Mumble

2020-07-06

2020-06-29 - Party Time

2002-06-22 - No party

2002-06-15 - Merge party

Tooling MRs

List of MRs for tech writers

2002-06-08 - No party

2020-05-25 - Merge party

2020-05-25 - No Party

2020-05-18 — Merge Party

2020-05-11

2020-05-04 - Merge Party

From #devteam:

2020-04-27

2020-04-20

2020-04-14

2020-04-06 - Merges & Beers

2020-03-30

  • merge party next week and all weeks from 3 to 5 PM CEST
  • status of the release
    • Docker
      • adapt alphanet.sh to the multinetwork node by adding –network $name-of-the-script
    • Teztool
      • not for this release, should be announced first
    • opam install tezos-client
    • (small discussion around !1689 : opam dependencies cleanup (or the minimized !1695)
    • merge the tiny vendored libraries (bip39, …) into lib_crypto or lib_client_base
  • coq-of-caml changes

2020-03-23

2020-03-16

  • soft freeze (no extra dependencies!) until we release a multi-network version ?
    • missing pieces for that are
      • externalizing resto
      • sorting out stuff aroung genesis of the different protocols
    • keep all the environment updates in proto-proposal
      • mostly the rust stuff
    • how are we dealing with adding rust dependencies in future releases?
    • Release of binaries / Debian packaging
      • may require 4.08.1 since it's currently the supported version by Debian/Ubuntu
      • build a static image with all dependencies
      • piping a shell script into bash will only happen over Benjamin's dead body
      • have a version that builds without the ledger library
  • ocaml 4.09.0 points
  • how are we dealing with old functionality for v0
  • naming around v0 and v1
  • proto_alpha: delete or merge Carthage?
  • next merge party on the 23/03/2020 at 17:00 - 19:00?

2020-03-09

  • it's merge party time
  • we merged 15 MRs

2020-03-02

  • Update on environment V2
    • Upgrade to ocaml.4.10.0 (or ocaml.4.09.0) (or ocaml.4.08.0)
    • Can't use Lwt.5 because of a long story that Pierre will tell you over a drink
    • Someone needs to work on updating Lwt to Lwt.5.0.0
    • Maybe update libraries
  • We should try to put the ledger behind a remote signer that does the queuing properly (?).
    • check the nonce revelation throughput of the ledger
  • Merge party next week
  • Paris Blockchain Week?
  • Carthage in 3 days? better estimate?
    • do enough nodes have the protocol to serve the sources over the network
      • all nodes already have the protocol due to the testchain
      • try migrating carthage on a testnet
    • should activate Thursday morning Paris time

2020-02-24

  • Seb's notes
  • We had a merge party instead and merged 20+ MRs that day
    • forgot to take notes though

2020-02-17

  • #694 → what do people think?
    • If the goal is to select the genesis protocol, do we need --genesis?
    • the idea is sound and should be implemented
  • Are we switching to Dune ≥ 2.x? is it planned? (heard there are already some upgrading problems)
    • currently on 1.11.4
    • combine with the migration to ocaml.4.08.1
    • switch to the last OCaml version that is supported by all our dependencies
  • Upgrade to last version of ocamlformat.0.14.0
    • or fix make build-dev-deps
  • Update on environment V2
    • merge a v2 (without changes) into master
    • then rebase proto-proposal on top of master and use v2
    • make sure that tezos-rust is informed and know about v1 and v2 and v3
  • Merbocop: current status → all tezos/tezos (∀ MR, see “Message to OP”)
  • Suggestion for next week
    • doesn't require more discussion, but more merging

2020-02-11

  • Backporting 006 baker to old context
    • doing the backport this week and Cryptium reviewing it thoroughly
    • after 006 activation announce that 007 will be an environment upgrade and won't support lmdb
    • bakers that don't want to upgrade should but they can still use mainnet-lmdb for the old context
  • how to grow the merge team
    • Mehdi
    • not necessarily merge access when joining
    • possibly in the next 3-4 weeks from CL
    • TQ: Seb will ask, some people on documentation/michelson
    • maybe ask Francois Maurel
    • Jun has no capacity right now
  • Merbocop report simplification (yes or no? if yes, who implements them? I (Romain) can, but Sébastien seems to want to refactor stuff first)
    • team is in favour of yes
    • Seb working on it
  • Unignore dune-project files !1453 => any reason not to?
    • MR patch is good, but MR itself needs to be fixed (title, description, resolve threads)
  • Remove nginx conf file !1509 => any reason not to?
    • We discussed at some point to have an easy load-balancer/protection layer for the default RPCs using nginx
  • Environment V2
    • https://gitlab.com/cryptiumlabs/tezos/-/merge_requests/39
    • needs review
    • give it names instead of v1 and v2
    • shell implementors have to implement a all environments
    • split the commits into what goes into tezos/tezos and proto-proposal
    • then merge v2 into proto-proposal and use v2 as a base for all env changes
    • then add "committing to receipts" as the first v2 upgrade in order to test it

2020-02-03

2020-01-27

  • 006 baker will require new context format
    • anything else that we need to communicate to operators?
      • higher CPU/RAM requirements?
      • what are our recommended specs for the node?
        • 8gb ram, 2vcpus, 80gb ssd for fullnodes
    • baking slack, twitter, reddit, agora - I'll loop Awa and Pierre/Marco
    • release communication about having to upgrade to irmin2 next Monday
      • with a backport for lmdb
  • environment upgrade
    • have a working version where the shell can contain two different environments and select it correctly based on the protocol declaration
    • all upgrades are merged into V1 for now and then we snapshot it into V2 and revert V1 to the old state
    • Sylvain to connect with Benjamin for extra info
  • operations receipt are on a branch and need review (not yet in env V2)
  • updates on the endorsement missing problem
  • simple logs (https://gitlab.com/tezos/tezos/merge_requests/1529, started by Romain and then Pietro experimented on it): interesting direction or not?
    • have simple logs that are less overhead to use
  • should we include a changelog in the repository tree (https://gitlab.com/tezos/tezos/merge_requests/1321)?
  • plebia updates
    • trying different caching strategies
    • building tooling for the better testing
    • slightly smaller disk footprint than irmin2

2020-01-20

2020-01-13

  • tezos-protocol-compiler: Seb → TODO 1st MR with just the code
  • how to handle security bugs and how to discuss, fix and release them
    • have proper channel and simple process
  • Hashtrees
  • vbot has a new store that is much faster
    • saves disk space / ram by a factor of 4
  • client can typecheck Michelson without node

2020-01-06

  • 83 open MRs
  • mulit-network node

2019-12-17

  • Multinetwork-node
    • Pierre is okay from his side
    • Romain can decide what to do next
    • Seb is merging it and 006 is added into it
  • Could someone merge the RPC docs
  • Please add issues for next version of the bot: https://gitlab.com/smondet/merbocop/issues
  • Anything that we need to push before the end of 2019
    • Mainnet released
    • Carthage released
    • Emmy+ blog post
    • We are good
  • Call next week and until the end of 2019
    • 30th of December is the last call of the year
  • Features to be merged in early 2020
    • Mempool
      • still needs a bit of troubleshooting
        • memory leak that resets at 22:00
        • some problems with some operations
    • New store

2019-12-09

2019-12-03

  • mainnet-staging → mainnet status
    • mainnet-staging release today (from romain@mainnet-staging-opam), then to mainnet in one week
    • romain@mainnet-lmdb → mainnet-legacy(-lmdb)
    • → release-page + blog post with changelog (with blob from Tarides?) → @VBot
  • escape-the-functor issue + voting in testnets
    • white-list of protocols
    • improve protocol compiler to avoid escape-the-functor
  • mempool status
    • related to multi-network-node: stuff from mainnet to backport, incl. “mempool-filters”
    • Ben: 10% of review
  • multi-network node: next release
    • switch to release versions (& betas, alphas, etc.)
  • Merbocop running since Friday, with --only-updated-in-the-past 2.5:days for the MR comments.

2019-11-25

  • status of the next 006 release
    • Assert Fix - needs review (vbot is the approver klakplok is the merger (or the reverse))
    • Emmy+ optimisations - status update (moving forward! let's see at the end of the week where we are)
  • mainnet-staging becomes mainnet status
    • Merge last irmin-pack (with "incremental merges") in master (pirbo is the merger)
    • Start a branch release@mainnet-staging with "alpine 10", last irmin-pack and "storage upgrade to 0.0.4". Add storage reconstruct and storage recovery once they are merge into master. (romain is the scheduler)
    • "storage reconstruct" review done (vbot and pirbo are the tester on mainnet) klakplok is the merger
    • "storage recovery" (klakplok is the approver, pirbo is the merger)
  • How do we organize with respect to externalising libraries?
    • externalized library should only concern stuff we have reasonnable assumption that
      1/ we won't have to refactor them in the near futur
      2/ someone outside tezos ecosystem will use it
    • General issues:
      • What about hotfixes as upgrades of our opam infra are heavy: OK because in last resort we would vendor back.
      • Who reviews? Not clear is the merge team will also take the lead there but for sure having both code released and still in the repo is the worse and has to be avoid asap(by actually merging stuff in the external repos and by using that in tezos)
    • Specific issues:
      • we need an informal communication channel to ocplib-json-typed which is central. We'll try one last option but if it is not working we'll fork.
      • tezos-crypto is depending on external but vendered crypto libraries which aren't really in a releasable state but that must be released to release… (HaCL release should be the definitive answer, there is no ETA for HaCL release :-))
    • tezos/ Vs nomadic-lab/ organization? (not tackled)
  • merbocop update

2019-11-18

  • discussion around contract storage assert
    • go through asserts and check them
  • discussion around Emmy
    • patches will come from Benjamin
  • discussion around how to deal with the reset of carthagenet
    • go back to 005 and vote on it again
  • fix mempool behaviour that when it dies the mempool restarts
  • storage2 works on every cloud except AWS
    • more investigation needed
  • remove legacy events from shell
  • merge new mempool

2019-11-13

  • Services, errors, log-events in separate libraries? for example _formats?
    • -> move into services
  • Making proto_alpha able to switch from babylon (with !1359)
    • /assume multinode
    • step 1 snapshot 6 into 7 in master
    • add a patch to unplug migration (5 -> 6) and plug a basic one (6 -> 7)
    • start proto-proposal
    • all patches on 006 must be replicated in 7 in MRs to master
    • proto-proposal can be rebased EZly
  • Where are we about the hunt of inefficiency in the bootstrap pipeline?
    • not the priority anymore. Priorities are:
      • reconstruct the context from a snapshot.
      • a patch to recover from context hash not found which would try to restart from the predecessor of the "head"
      • Logging mustn't allocate if it wont emit the message in the end (because it is debug message and we are not printing at level debug)
      • A patch to make smaller steps in bootstrap pipeline
  • Are we testing the last development in irmin (-index and -pack)?
    • We'll sort that out around a beer dans 1h30
  • Anybody at NL responsible for trying ocaml-memprof and ppx-bisect?

2019-11-04

  • Shell release in mid November
    • potentially include the new storage backend as a mandatory upgrade
    • Problems left:
      • Behaviour of merging
      • OOM issue - seems fixed
  • Merbocop startegy? Merge everything + tezos MR? or somebody wants to look in there first?
    • Outcome: Add it directly to tezos/tezos
  • Multi-network node:
    • Seb: I've started using it in the flextesa docker image.
    • What is left to do?
  • "assign" button policy
  • MRs for 006
    • Baking rewards
    • Comparable pairs
    • Baking RPC
Last changed by 
 
Adrian Brink
0
2193

Read more

Cryptium Labs - Kusama Validator Addresses

Current - [Update 2020-06-10] Cryptium Labs is currently running two clusters of 8 Kusama validators each. The recommended strategy for all our delegators is to take their KSM holdings and evenly divide it across two addresses. Delegators should nominate the first cluster from their first address and the second cluster from their second address. This is the most optimal economic strategy for Kusama and eventually Polkadot. Should we receive enough delegations to both clusters, we will add a third cluster and inform all our delegators. Below you can find the two clusters: Kusama Cluster 1 GcqKn3HHodwcFc3Pg3Evcbc43m7qJNMiMv744e5WMSS7TGn Fy2rsYCoowQBtuFXqLE65ehAY9T6KWcGiNCQAyPDCkfpm4s Faaf2qgwtaFum78U8TdEPnUd4u7HPk65cdaFj7KbEA6wi2Y FPRN9EfcCTswkRXfrkwoouep8uhVJRvMy1hnhB4Hb73biwK

Jun 10, 2020
Cryptium Labs - Polkadot Validator Addresses

Cryptium Labs is currently running one cluster of 8 Polkadot validators. Since the optimal strategy (the one maximising returns DOT returns) is to have as many validation nodes as possible, we may add more nodes to cluster 1 and/or create more clusters. When we create more clusters, we ask our delegators to nominate all new addresses/clusters. If you delegate to us for the first time, please send us a short message at polkadot@cryptium.ch . Thank you and happy nominating. :-) Cluster 1 12C9U6zSSoZ6pgwR2ksFyBLgQH6v7dkqqPCRyHceoP8MJRo2

Jun 10, 2020
Carthage 006 Changelog

Nov 12, 2019
Read more from Adrian Brink
Published on HackMD