Everything EOS EOS Leap Upgrade Group, 20 Jul 2022

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: EOS Mandel Weekly Meeting #24 (July 20th 2022)

Description

This week the huddle group discusses the release date of Mandel RC3, testing on Jungle 4, FILLPG feedback and more!

Timestamps
00:00:00 Recap from last week and today's topics by Daniel
00:00:48 Setting objectives for the consensus upgrade by Daniel, Yves
00:04:04 RC3 update by Stephen, Yves, Ted
00:07:04 D.U.N.E update by Stephen
00:08:00 Testing on Jungle 4 by Michael, Yves, Corvin,
00:19:45 FILLPG feedback by Randall, Michael, Daniel
00:24:02 Kylin update by Daniel, Denis, Michael, Yves
00:30:28 End

🟢 In progress action items are on track:
Align on objectives #38
The target date for consensus feature activation was changed from September 23 to September 21 to avoid activating on a Friday.
Mandel software development #41
RC3 release was a no-go for this week. Targeting release next week.
It's possible RC4 will be required before we go to final
Next release (RC4 or Final) would likely follow on August 3 or August 10
EOS Support has shared responses to the survey requesting feedback on fill-pg:
2979424562146452305fill-pg-feedback_update_20july22.pdf

Platform Support #55
D.U.N.E. release did not make it out last Friday, targeting this Friday
Test Mandel on Jungle 4 #44
Consensus features are activated on Jungle4 🥳
Hyperion is working 🥳
dfuse stopped working upon activation, but a fix is underway. dfuse will be needed for UX's IBC solution
Need to look into activating BLOCKCHAIN_PARAMETERS consensus feature, which has not yet been activated
Testnet strategy #40
EOS Nation and EOS USA are working on getting Kylin prepared for testing
Participants (22)
Host: Daniel Keyes | EOS Nation
Yves La Rose
Matt Witherspoon
Kevin Heifner
Stephen Diesel
Bucky Kittinger
Aleksandr Aksenov
Denis Carriere
Alysha Mims
Ted Cahall
EOSUSA Michael
Jeff Werner
Corvin (cmadh) | Liquiid.io (Corvin Meyer auf der Heide)
Ross (EOSphere)
Brendon Ross
junhank | NodeONE
Max_Cho EdenOnEOS Member-South Korea
Vlad
Randall Roland
Genie Kim - EOSeoul# NEOPLY
g
Shaq (g)

Everything EOS is the official media channel of the EOS Network Foundation (ENF). The EOS Network Foundation coordinates financial and non-financial support to encourage the growth and development of the EOS Network. We’re harnessing the power of decentralization to chart a coordinated future for the EOS Network as a force for positive global change.

// EOS Network Foundation //
https://www.eosnetework.com
https://twitter.com/EosNFoundation
https://www.linkedin.com/company/eos-network-foundation
https://t.me/EOSNetworkFoundation
https://discord.gg/eos-network
info@eosnetwork.com
#EOS #EOSIO #ENF

A

All right welcome everybody. It is july 20th. This is our 24th week meeting. uh Let's quickly recap: last week we checked in on progress towards rc3. um I think we're targeting a release this week. I believe that hadn't happened. We can check in on that in more details in a bit um platform support.

A

um I believe we we're expecting a release friday as well, so we can check in there I'm not sure what the status is um and testing on mandel. We do it with definitely lots of progress to report on our testing on on jungle. Floor features have been activated, so we can get more into that.

A

um We'll talk about our progress on kylin and I also maybe we'll just start with actually before we even go into that before I forget. I wanted to talk about the previous, so we had set an objective uh to activate the consensus features on september 23rd um september. 23Rd is a friday, though, so I think we'd like to instead move that a few days up to a wednesday, so that would bring us to the 21st.

B

And the only reason for the 23rd is the uh the last consensus upgrade was on the 23rd, uh and so it was somewhat symbolic. That was the only reason for it, uh but yeah instead of having it on the friday, we seem to be releasing on wednesdays, which is a little bit better, gives us a little bit of time at the beginning of the week more time afterwards. In case something happens, um and it's still pretty much the same.

B

It doesn't really change the target.

A

No shouldn't make an impact, uh I don't think anyone's gonna notice, the oh, you didn't, you didn't upgrade on the same day as as 1.8. What's up with that, uh I think that's a good good, uh good idea. Let's, let's make it on the wednesday um we are working on an article to kind of spread. The word hey get ready for september 21st.

A

The activation is coming, make sure you're you're prepared, so that you don't you don't go out of consensus with the network if you're running any nodes yourself um so expect some.

A

uh We already got a draft of that we'll we'll probably be reaching out to some folks to get some feedback, make sure we're all on the same page. uh Maybe we'll share a draft before we publish it and um and we'll work with the eos support team to you know: they've collected already a big database of dapps and uh exchanges so we'll we'll leverage that to spread the word.

B

Do you have that list by chance, or are they the only ones with that list.

A

I've got I've, got it too. They've shared access with me, so they've got it in their uh yeah. They've got a system with all that in there that we also have access to um so yeah. That's going to be super helpful for these communications in an ongoing basis. Now we've got, we've got the database.

A

um That kind of all that the hard legwork of collecting those those contacts has been done.

B

Can you share the article uh draft prior to uh publishing it so that we can see.

A

Yes, absolutely yeah! Thank you! So I'm I've already received a draft. I haven't had a chance to review it yet myself, but I'm going to be uh working on that in the next day or two and um and then I'll share it share it with everyone else. When it's ready.

B

So mandel release- uh I guess final candidate when it's out that will be the first time that the community actually has a a full release, full-fledged release that wasn't created by block one or released by one that everybody would be on boarding to. But once that hard fork happens then not only uh would be essentially forking out block one we'd be forking out bullish working out voice, so they would need to onboard onto ours if ever they wanted to continue down that path. So that's quite symbolic as well right.

B

It really completes that that breakaway from uh block one.

A

Yeah yeah exciting times, looking forward to that um and yeah. Maybe we can get an update on the so we were expecting rc3 this week. I I heard that's not the case. uh We get an update from someone on the e f side, yeah.

B

Oh I'll hop, okay, I'll go for it. Stephen.

C

Oh sorry uh yeah, so the the the update is that uh it's a no-go for this week um optimistically targeting next week, but we'll track as things progress. uh Basically, I think that there's just a couple overlapping issues related to uh modular exponentiation, just we're trying to set appropriate limits on that, but we're also working on removing a gmp dependency that overlaps with it. So once we get those two interlocking items wrapped up, we should be good to move forward with the rc3 optimistically targeting next week, but uh we'll keep you posted.

C

um I think that that that's really all to report on this front unless you want to do an update, but I know that's kind of the next section.

A

Sure yeah, let's um save that for the next section, anything else, though, on on rc3 or mandela.

B

Kind of so I guess if we, if we are now releasing rc3 on the 27th, the target for final release uh was august 3rd, which is still possible, but that would only get one week. So it's probably more realistic that final release candidate would release rc3 would become final release or by the 10th. We'd have an answer type thing so either the 3rd of august or the 10th of august.

D

Yeah, it is also possible. We won't put rc3 out next week. It's only tentative and there's enough complexity in there that it may not go out. So you can see we're we're kind of slipping a little bit here and it might be. You know we. We won't be sure that rc3 itself will fall into rc final. There is some interesting code in there, but uh we will be keeping you posted and those meetings are getting much more detail and thank you, stephen for running in such great detail.

D

This uh last week, we've really upped the game there with steven running the uh all the dependency charts and everything through uh the github project management. Now.

A

um So it sounds like so there may be an rc4 given the complexity of changes in rc3, so either we'll have the final okay.

D

Cool I want to make that clear that that is possible, we're not actually planning for it, but obviously, if we find stuff and once our c3 is out, we might have to make an iteration and test it, and it is enough complexity and the two things were the two pieces that stephen talked about, that there could be a bug in it.

A

Great okay, good! Thank you for the update anything else to add on the mandel software release.

A

uh In that case, why don't we move on to dune.

C

Yep so uh it's kind of the same update as last week, but a week further along, so I think we're shooting for thursday or friday this week, there's just a couple of outstanding issues, but I know that uh bucky and lynn are addressing them feeling good about this friday. um This, of course, is just a lower risk item than the other stuff that we're tracking at the moment, but that is the latest that we've got.

A

A

D

A

um Anything else on dune or we, otherwise we can move on to jungle, testing.

A

All right so uh features are activated on jungle. I believe all of them, maybe not all of them. I think I saw a note from matthew that perhaps there was one that we did not activate.

A

I don't know if anyone can confirm that.

A

E

A

E

Was it I saw matthew say it, but I don't know, I don't.

E

Yeah he was looking at a summary of all the protocol features and it said it's blank and that's what he was picking up on. So all the rest of them were set, but not that one.

A

Right, which I believe we can see that.

E

I think that may be on the jungle blocks if it's got jungle four, because I think that may be where he pulled all that from.

A

I think he got it from our api here. Let me just make sure I can check it here.

E

Oh, maybe it's a chain info on the validator. You know.

A

Yeah, I'm trying to pull it up now. How do I get to the 1 4 jungle? That's the question.

A

uh Anyways the features are active. I don't know if anyone has anything to report on how their nodes have responded. History is one thing we weren't particularly interested in and hearing how that goes. um Diffuse is down. So that's that's not working uh any any luck with hyperion, michael or anybody else.

E

Yeah, so when I shifted over to rc2, it picked a yeah rc2 had the ship fixed and the original reader, which is the 1.8 version, picked up on the the ship and started indexing going forward still says, it's happy seems to be solid.

E

um As for when the protocols activated pretty much all the nodes rc2 carried over except I did have one that wasn't a bp that did have, I think the original rc1 protocol and it forked, and so this guy just stopped it deleted. The protocol features folder. It created it whenever it fired it back up with the right feature, id and immediately sync back up to the chain so um solid.

E

But if you happen to have rc1 yeah that that little uh feature grid change did fork it out, but it was very easy to recover and- and history seems solid- I I don't really use it much, but the index is running and I think we're feeding, jungles or blocks so box.

A

Blocks is still working is.

B

That is that one upgrading from rc1 to rc2 only when making that particular upgrade.

E

I had some nodes that were rc1 to rc2. I had some nodes that I did, I think, take from 2013 up to rc two.

E

I had been holding them off until I tested it with a new server and then upgraded the other server. So all of them are pretty solid. How to use snapshots, you know or a replay to to get them back. You know the state back in sync, with the new build versions and whatnot uh the sealings and what but yeah it was.

E

It was pretty solid and like say the only blip I had, um which I knew I had one or two out there was that they, the few that had that that original rc one grid uh feature the avm feature and it.

D

Definitely did pieces.

E

Yeah yep, it just caused it to fork out and like say I stopped it deleted the folder and it recreated all the versions it wanted, and then it immediately synced right back up to the chain. But- um and that happened to be a ship node, it was not a ship node that I was using. It was one that I had tested with. So all it all seemed pretty solid. As long as you didn't get into that rc-1 hyperion survived through the protocol activation.

E

I I I can do some queries to make sure, but.

B

And daniel on your side, matthew or denise here so diffuse is not working or it is working.

A

Diffuse is not working. Definitely not so. We've got those chain, those fixes coming in rc3 that, I think, will perhaps we'll see if that, if it works once once those get rolled out um and there's some work.

B

Was working in rc1 not working in rc2.

A

Well, not, we never activated the features but and we were expecting that the feature activates. So it's the activation of the features that caused the fuse to stop working, um which we hadn't we hadn't tested with rc1, but we were expecting that this would be the case.

B

So, for those fixes do we know if uh stephen watanabe or todd fleming are the ones fixing that, because that was part of oh yeah,.

C

Oh sorry, I was just waiting for a break in the action there. Yes, so I I know that peter oswald had resolved the one issue that had been submitted. I think that there's been testing against the branch there to verify that it was successful, and I do believe that there's been further testing to ensure that there aren't additional bugs beyond that.

C

I would still revert back to the latest update from matthew darwin just to make sure that there's nothing new, but there's no newly submitted bugs in the repo to observe any additional problems uh with diffuse.

A

Yeah, I believe the initials so we've been testing sort of that already it's not released, we've got, we've got access to it and from what I understand seems to have helped whatever whatever you guys have been working on there.

B

So my question is who's been making those fixes. Is it the fractality? It's just.

C

It is not the fractality team, it's the.

A

C

B

Is the fracturing team should be the one fixing this so we paid for this, so I don't want to double pay for oci and for fractally to work on the same thing that we already paid for. So that's my just my concern. It should be the fracking team that should be fixing this.

C

Sounds good yeah I'm! I I'm happy to discuss that with you um outside of this meeting. Yeah.

D

When, when he just talked to me about stephen, I know enough enough about it, we can I'll fix it.

E

And just to bring it up, it does look like on the validator that hyperion's got active actions. I got all greens on it, so I must say it looks pretty solid so far.

E

Sweet that rc2 fix is what the the one fork was for the reader.

A

And uh this is the protocol features activated currently on jungle. Four again, this is the one that we were saying looks like has not been activated yet. So this is one of the new ones right, one of the ones that came out for mandel.

A

So I guess that's one we'll have to revisit make sure we activate that get it tested, but the ones that you guys are really interested in getting activated are the primitives. So we can start begin. The um and evm testing is that right uh correct?

A

So we should have the features.

D

We need activated for that.

A

Yeah, so that's that should be good to begin any any testing you need there.

A

um So yeah features are activated.

A

Hyperion's working.

A

We need to look into.

B

And one of the reasons why diffuse functionality right now is important, um and so I just invited yelm and shaq on the line is because it is needed for ibc, which we are also deploying.

B

um It is the method with which we'll we'll be able to communicate, so we need to have diffuse working and it was working. So we really need to. uh That would be a showstopper.

A

Yeah, so so weird um yeah there's work to do on the software side downstream of something there's the work that fractally team did. Then there was work that actually has been done on the defuse software to be compatible with the with that that work is underway um and yeah like, like, I said, the seems that the the work that has been done on the mandel site is helping, but there's still yeah lots of testing and work on the defuse software side. That's that our team's working on right now.

D

And um this is um guillaume speaking. um If there's anything, we can help with happy to uh sync up with your engineers and.

A

Awesome that's great to hear thank you.

D

Thank you again. This is needed for for ibc.

B

For the generation.

D

Okay, interesting.

B

All right corbin, maybe if you want to touch base on that feel free, I'm guessing y'all, will will gladly share. So we have a different ibc calls for this either you could, you probably join, I believe we record and publish the calls or you could reach out to uh or shaq.

D

That that that would be great, especially because we are working on yeah, a diffused light version, not sure if you know uh so, if possible, it would be great if our solution is uh compatible with what's needed from the fuse for ibc.

D

I assume I will ping you yeah just uh this is shot here uh just to let you know, like the part of the fuse we're using is just the fire hose service, so it only relies on uh like merge blocks in a sense, so you don't get all the bells and whistles of the rest of the architecture.

A

Fire hose is much lighter weight. um Yeah we're we're looking forward to moving to a future where fire hose layered on with substreams, rather than the bulky bulky defuse, is going to suit the needs for everybody, but that's a work in progress in the meantime, we're we're getting diffused, making sure that's working. um I guess just.

B

Good to know they're fine well so to clarify it's not necessarily that the fuse or fire hose is needed for ibc. It's that that's what the ux team has been using to fetch the information they're look they're they're using, but on the roadmap, there's going to be other possibilities, but yes in the meantime, for the short term, that is the means by which they they uh they do obtain that information.

A

Okay, I'll check in with matthew and the team on the status of fire hose versus defuse and whether there's a difference in effort to make sure they're working well, we'll prioritize fire hose.

A

If, if there is indeed a difference, um okay, anyone else have anything to report on testing with features activated or the process. You know how is the experience of upgrading anything. We need to any feedback, for maybe the the enf on the documentation.

B

Or anything like that, randall posted something about five minutes or so ago in the chat randall. Is that something that you want to bring up or is that was that in regards to something we were talking about.

F

Oh yeah yeah that has to do with uh phil pg, um so we were able to receive um some new response from some developers and I think one developer. One responded said that they've never heard of postgre. They prefer mysql.

F

uh We have another respondent say that the open source use io explorer uses field pg. um Several networks also use the tool, um and you can also use it for local development.

F

Then uh last respondent um stated that um um they are not using it in their project currently, but they are working on um integrating post gray, in addition to mysql tidb and elastic search as a storage backend into a diffuse light solution.

F

um So what what else? You said that it's likely that using their solution using their um solution developers will be able to replace post grade instances followed by the fill pg plugin. um So I basically just posted that in the zoom in text, so everyone can read it. It makes uh uh uh if it makes any sense and uh it helps with uh whether we keep uh filled pg or we um stop. uh I'm supporting it.

B

Sorry, what was it, what was the context.

A

So this was, uh there was a an ask from I believe it was kevin was uh wondering if phil pg is still something that the community of developers are relying on. There's a, I guess the potential of stop supporting this in a future version of mandel, and I guess I guess the we don't. It doesn't look like there's a lot of a lot of people using it. So we wanted to confirm that.

B

Gotcha, thank you. Thank you randall. Thank you. Sorry. Are you capturing that I guess yeah you're gonna capture that in there.

A

Yeah, thank you. These are all cool. We've got a few respondents here, so that's all documented in this issue here for kevin and anyone else who's um who's interested. Thank you, randall. Sorry, for uh I missed your note there. I skipped ahead.

F

No, it's okay, no problem.

A

uh Any any questions or comments on that from anyone.

E

I mean that they're saying what we kind of figured is that yeah there are some people out there that use it. I'm not sure what block explorer did block one release the open source block explorer. They use yeah.

D

It's the one you have in the esio and the old repos.

E

Yeah, so it's just the old test net or whatever or the old repos anyway, so I mean that's the problem. Is it's going to be scattered throughout random people that used it? But I don't. I don't know whether that answers keep it or not, probably more towards not there's only three responses.

A

Yeah yeah yeah, the question was, I think, was it one question on the survey correct me if I'm wrong randall, something like do you currently or have any plans to use phil pg in any of your projects? Basically, that was the that's the question.

E

Yeah, that's correct. One said no one said: yeah we'd love to pump it into a database, but not postgres and the other one said yeah, a bunch of people all over the place. Use it so or could use it, not that they do use it that, but those are the ones the private people you just don't know.

A

Right right, okay, well, some some insight for the team um back to testing on jungle. I think we covered it unless there's anything to add.

A

In that case, we can move on to. um I don't have much to report on kailin. We again, we've got our um we've got god mode now we're sharing with with uh eos usa, michael, um so denny is going to be coordinating offline with with michael to get what needs to be done um here.

A

Basically, we need to get power up going. We need to do binary, updates eventually we'll get to future activation. We need a faucet uh anything to report on that denny. I see I think, denny's on the line.

A

I think you're going to be working on the faucet.

D

Yeah, nothing yet to report on that, but I'll sync up with uh michael there, oh and we'll try to hammer a few of those things.

A

Cool is there anyone on the on the line, who has any sense of what is most needed? I think we want to start with the faucet right.

D

Yeah we can, we can talk about it. Yeah.

E

I'll say that seems to be, I mean, there's still random people to come in there and either need an account or need some tokens. I mean we can manually handle it. I mean we gotta, get the network a little more stable, either wake some people up or kick some people out and slide some more people in to at least get things going, but I think that's kind of the next.

E

Those would be the two things infrastructure and then feature dev, wise, probably the faucet, so people can use it.

A

Yeah and I think, there's some features pre-mandel that are not yet activated, so we need to get those features activated so that it matches what we currently have on mainnet, so that we can then test to the upgrade. Is that right.

E

Probably yeah I mean we need to get them in sync,.

B

Well, now that you have god mode, wouldn't you be able to do similar to what botan does on jungle and somewhat force the update.

E

uh Yes, that's what I'm just trying to gauge a bearing on how big of a shotgun we use for it. But yes, we can that's what I think kind of starting to put the foot down on the gas on that and and asking and then just doing yeah a couple days.

A

There's some work to do to figure out exactly what needs to be done, I think is the first is: is hey? What are, what are the deltas? What needs to be activated? What are the, how far behind is, is kylin exactly um so. That's part of that work is the idea of michael.

B

Because I know, michael denis, you guys run multiple nodes with the idea. If you can't get the buy-in quick enough, you'll just slot yourselves in for 15 21 minimum activate everything and anybody else then who wants to come in has to upgrade in order to then be put in.

E

Correct correct, and ultimately I mean we have enough to have consensus, it's whether we're okay, with a couple of block producers being out of the yes. We know we're going to be three short even before we activate, but.

D

That's right, though,.

E

And then that's what I was but.

B

That also tests, what happened once that happens, or if that.

E

Were to happen.

B

E

Why I was like ah it's an intern, of course, we've missed a lot on jungle. I mean we've actually had some some good scenarios. So anyway, that's what I was thinking about that is, uh I think, we'll just ask one more time and then yes slide a few more puppets in there and and just start banging things through and then denny can get the list of all the things you know.

E

I think he always did those protocol activation or help get a bearing on it from that aspect, I'll tackle the technical and spin up a bunch of minions awesome. Thank you, michael.

A

Thank you and and sorry back to jungle. Besides looking into that blockchain parameters, feature. Is there anything else that we want to focus on this week between this week, and our next call uh looks like we're pretty happy with with hyperion I'll, have I'll get some more answers on the state of defuse for our next call, um but is there anything else, for example,.

E

Not so much on the evm, I may try a few other networks, because there's been a few weird reports of quirks that shouldn't have happened but like forking of nodes, after an update, so I made not so much of a jungle. Obviously they're all yeah.

A

I saw that on telos.

E

I may I may spin up a node with mandel and just you know, fire it off on all the chains and see how it handles it just to see if there's any kind of quirks, but um that's it as for jungle, probably not.

A

Yeah we've been running mandela on a on a I think. Most of the chains we've been we've already got it on main net on all of our api. I know it's just on our bp node. Obviously, there's not features activated there, so there's different different things to be tested, but so yeah, our diffuse, is working fine on mainnet, with the features not activated it's with.

A

uh Once we activate the features that it becomes an issue uh okay, so we will continue to monitor how things are going in jungle, we'll double check into this feature and get it activated if it needs to be and I'll get some info on jungle and and uh yeah, like I said, demi and and michael can work on kailyn.

A

We talked about a little bit already. The change in the release date I'll have perhaps between now and next week. I'll also have a draft to share on the article that we're working on so something else to look forward to on the we've got an issue here for that communicate, upgrade plan. Broadly so andrew ware on our team has been working on that and a whole bunch of resources. Basically, assembling a bunch of resources on and summarizing kind of, the the headline get ready for september 21st, um so yeah I'll share that one.

A

That's ready anything else to discuss today.

A

All right well I'll, wrap it up and have a great day everybody. Thank you for joining.

E

D