Apache Mesos Containerization Working Group, 5 Oct 2017

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: GMT 2017-10-05 Containerization WG

Description

Agenda and Notes:
https://docs.google.com/document/d/1z55a7tLZFoRWVuUxz1FZwgxkHeugtc2nHR89skFXSpU/edit?usp=sharing

A

B

Hello, hey goober,.

A

B

Sorry I don't realize that I had to join the meeting before this meeting can started. Let me pink folks on the slack yeah I, just post like this mean to be auto, started.

A

I've been the news in support that we.

B

Had some support hinge as cast to join on first, okay,.

B

All right that's way for a few minutes.

B

B

Okay, so I think let's get started, so we have two gay guys hear me okay, so we have to agenda item today. So um cheetah and Gilbert is gonna. Talk about on the recent work.

B

Garbage collection on container image layers in mixes continue Iser and then I'm gonna have a discussion on the VM support. I. Think someone had this note like that. This agenda item here: okay, let's go with the the garbage GC image layer, stuff. First Gilbert: do you want to share the screen or hey.

A

You want to talk about that. Oh you want me to do it. I can.

C

Try to share the screen: okay,.

C

Can you see this yes,.

B

C

So this this is a so this work is, has been ongoing for a cup of month on and off so the goal here so so. The background is that the message continue. Editor is able to provision the long stalker containers from docker image, but right now there is no way to to enforce that's the disk space news. 8 of this is bounded and, and we could leave images the athletes, for example, in our GPU cluster, which metal containers the only twice. We occasionally see machine completely unusable because we fill up all the disks.

C

They are a great doctor images, so this kind of becomes a really high priority problems for anyone. That's operating vessels contriver introduction, so we modified a little bit right now. The goal is to just provide a working MVP so that we can wish there is a way for the operator to ensure the disk space of darker darker image. Star is bounded, and we also want to making sure that the implementation will you keep the vessel controller usable meaning we are not going to lock the start down for a very long time. That's needed!

C

Neither the darker pool neither of the so that other pools happening at the same time can still proceed in the.

C

So right now, right now this is the MVP we plan to implement a new message in an operator APA action, so we can just trigger this through through through a HTTP HTTP API to print any unused images. So, in terms of design, there are generally two at least two possible directions of implementing similar things. One is a marker sweep another would be a reference coming. So after a couple rounds, we decided to go with a marker sweet for now, because it seems seems simpler to implement.

C

It does not involve changing the internal data structure of the talker store. I'll cover why the reference counting seemed to be a little more complex on that in that fashion. First, let me explain how we currently and students so right now, the inter talker store is kind of a private component of the Container Iser and provisioner.

C

So so we realize we can simply rely on the continued writer to collect order to crack all the images that we need to retain during during an image provisioning, round-eyes, just Traverse, all the containers execute a container nesic dinner, or you will stand alone in the future that the Kenyan writer has simply travels all of them and collect what image they are using and then pass that down to the dock, respond to stock and then compare that this list with all the other image it has.

C

It is, is check, pointed image, set and determine what images are not needed anymore, and what layer that I need anymore. After that the occurs, the token stars, metadata manager will find a new. The new image Center is only images that actually needs to retain, and also after and after that, doctor star will move all the layers. Other I use the layers into a new direct in new temporary directory. This is the marking phase.

C

So the reason we do a move incident delete him instances because it will be deleting hundreds of gigabytes of data and that could take that could take quite some time, but if we simply move the layers into a temporary directory and then delete them asynchronously, we don't need to lock the star to ring out while we're deleting these stairs.

C

So the intern story is going to be the interest or has to be locked during this barking thing. But luckily this is pretty much reading a file. Do some calculation checkpoint pile?

C

They move a bunch of directories, so hopefully it should be every chip to a very cheap in production and doing this, the sweeping phase this during the spin phases or can already be unlocked and there one one broker to implement this- is that we realize that container config was not previously checkpoint, eight, as of as of one point pointing so that means, if the mass of agent recovers, for example, it restarts we actually don't know in. We actually don't know which image is used to launch a nested container.

C

This is this, is this is specific, especially problematic for and that's the container, because there is no previous checkpoints about how it was launched at all. So after well, agent restarts we, we don't know one if weather NASA container is using, which images using so there's no way we cleaner is pretty fun to safely implements.

A

So since it out so this is also affect the top-level container right.

D

C

I think the top-level container it is possible to replicate the logic of how it is whomped, although how it would like to try to replay how is launched but but yeah. If there is code change in missiles, agent or something else, then there is a chance that, whatever we think so, we or whatever we, whatever the current called thinks it. How it was launched, may be different from how it was actually launched.

A

So basically, we one day we want to continue config. Here we want to find out the mapping like, for example, given and continue ID. We would like to know what image this running container is using. So let's say you have a level continue and you restart the agent or you acquire the cluster, and if you don't have the container config, because after recovery we don't path, fill the container config for any power level continue. It means you cannot figure out like what exactly the image this running container is using.

A

So we have no way to do the GC, no matter for the for the top level or the nasty container after any recovery right.

C

I think for top-level container, because we info we can recover them indirectly through other information, although I don't know whether we will trust with the back ceiling logic wonder-percent. If the logic of methyls agent cause changes, we could have, we could have incorrect view of something, but for the nested container there is absolutely no way to actually leave information at all. Okay,.

B

So uh sorry, I I was going to stretch it a little bit so on kay guys, explain to me like what do you are doing right now to to bypass these problems so.

C

We are not bypassing this problem were simply so. You know current in so in the current patch we wrote we, the current cost had actually relies on this behavior, so this is actually implemented as a stepping stone to the image TC.

B

Rely on this behavior, what do you mean by this behavior rely.

C

On checkpointing recover I see.

B

So I think you're saying that if you have some legacy container, don't have container can check pointed the the reference counts. Are the system? Rockin sweet might not work because.

C

Yuning, we simply are all to say that is we don't consider it safe to do this so.

B

Okay, I see you air out when someone hit the API and say it's not safe to do that, because some containers don't have that information being checked. We.

C

Probably will make it clear to operators say they should either dream this agent. So all the connect, recycle all the containers or wait until all the containers to around okay.

B

Yeah I think that makes sense to me and you can add to the documentation saying that in order to use this feature you how to train H and reboot the age and probably or her don't need to reboot of agent. But you have to train the agent.

C

Drain agent, if they're not running very long running dots, they can simply wait out for all the containers. To finish.

C

B

You have to force that service to be restarted.

C

You know we operate cluster. We can probably wait for a couple of weeks until naturally called turn out. Okay, angling. We can deal with them separately. Okay,.

B

That makes sense a sense for me. I think another issue I have is: how do you guys still lead those layers? Are you gonna move to like a trash directory and then like Strato, the deletions that I don't like kind of like over clothing, too much I own, the discs that are affecting other tasks, yeah.

C

We move it to a trash directory, but we are not stretching yet. This erection will be necessary. Throttling.

B

I would say like a delete one by one, rather than just like fan out on to to to like well.

C

I think it even wonder what okay.

B

Yeah I just do that. An official Pony like restrict IO from that operation.

A

So G was your motivation compared to like delayed layer one by one compared to like the later oh geez, the trash directory? What's your.

B

Concern, oh my my cuz I, don't know how this arm Durr works. If you just like win file-by-file, just probably okay. My main worry the reason I'm asking this question is in my: if there's a lot of, are to delete that in my effect running containers in terms I, oh I, remember: I,.

C

Actually asks in the past whether we do any Isle nice on the yeah.

B

I, can you I own eyes on that I think.

C

That you told me we are not. We were not doing. We are not doing that right now, but.

B

We should do that is. Is that hard to do like having our own eyes? For that.

C

The sub process is easy, but if you I see.

B

Yeah, you have.

C

To be a subprocess Canadian, what our.

B

E

That is something that happens in in practice. Is that the GC for contain gc4 sandbox? Those takes up all the I, ops and kind of stay. It interferes with how fast you can stage and things out there. How fast you can stage tasks. Artifacts has.

C

Anywhere observed a serious problem because of this in practice,.

E

Serious problem, I, don't know, smoke tests will occasionally time occasionally blow their time budget, which is just purely arbitrary, and it causes some toil on our end. But I probably argue that that's self-inflicted made it big and but in general you know your stuff causes. This stuff consumes part of the limited resource of ions from your storage subsystem, yeah.

C

C

Yeah I think I think maybe using sub-process class ionized can can make leaving a lot of these problems.

B

Yeah but then you have to move the logic to a separate binary on which we might be able to do. I think, like the same thing, can be reused by by the sandbox GC.

C

B

Anyway, I think this night, probably not MVP, but we're think about how to do that so that I don't.

C

Actually, post MVP, nice they're sleepy right, we.

E

Can easily do I mean easily, but you could have a separate resource process which.

E

You could I was thinking about having you could have a super sauce. um This extent stands alongside the agent and you can just send blue processors to.

C

Do you isolate or calls into that? Oh yeah.

B

So we yeah I think there I remember, there's a to do there that we want to on to ionize sounded EU to so that we don't interfere with the well. It's just a limited, like kind of like, like reduce some of the interferences yeah I. Remember, there's a to do that using I'm nice. What do you look for? Do I? Think it's easier because she's just do ionized EU and you just want to make sure I own Isis. There.

F

So I have a question, so all the DC stuff is specific to doctor store so for things like oh yeah, so for OCA image, also I'll be using the doctor store only I'm just wondering if he so what happens when we we start using OC our we start supporting us here so so.

A

Basically so basically, this is basic. Gc supports our current Tokra store and in the future we will have a separate store, called unified, artifact store and the gist the OCI image. And if you depends on that unified artifact store and then we will happen, a new GC in with the OCI and then which make it compatible with docker image version. 2 schema right now we are working darker version, 2 schema 1. So this is a very basics possible key. So.

F

Buddy I mean my basic question: would he have to change a lot to make the same GC work with a different tool? I mean.

B

F

Oci image pic is coming great yeah.

B

So actually, this is the purpose. So that's the reason we want to move to like, like a unify, artifact store for not just artifacts from fetcher and also like layers from talker images, OCI images, whatever images layers so make a unify, and then we have a single GC mechanics and in a single cache mechanism like cache replacement mechanism for that cache store. Basically, so.

F

I guess what I was trying to ask is: would the current GC make a bloom change a lot 1mi.

B

We haven't designed that yet this okay, that's! This is like more like a stopgap.

F

B

Okay for production usage.

C

This is the primary blocker for us to roll out vessels controller at scale, okay, pretty much limiting this or a GPO cast only because we don't have other choice, but we are not bringing this sort of service class right yet primary because of this okay.

B

Yeah I think this is like a more stopgap solution, but I think it's good enough, and as long as you have a monitor script to monitor the entire host disk and if it goes low can just trigger that endpoint to garbage collect those I use later. That should be good enough for at least for MVP I would say the.

F

Other question I had was the the trash folder that you're gonna use right to actually move the GC files instead of deleting them. The trash folder has to really be relatively big as well. So would there be some I mean I, don't know whether it's just a documentation issue where you need to document that the trash needs to be it shouldn't be sitting in something like slash temp right, because otherwise it might happen that you just overflow, slash temp itself.

C

Is so I think the the patch has.

C

Directory is on the same level, said the doctor Scott. So it's probably.

B

Okay, you have to be in the same device, otherwise OS rename is gonna fail because OS rename only works atomically. If you are on a same device, it will complain if you move to attempt like slash temp, it's a different file system, even.

C

If it does not fail, it's way too expensive. So if you can support open someone as weak, if we move my to move the different device which we need to yeah, okay,.

C

So I'll proceed here, so we so to implement this. We actually need to add a new virtual virtual interface called pool tada to all the store interfaces, although in the MVP already the talker star will implement this. For now for the for app see, I think the only other store interested right now exact see. So that's what simply return. Nothing immediately.

B

C

B

Is how can hold on little bit so what's the actor, the images.

C

Active I think I think we actually want to rename the activity to rate ten images to make it clearer to make it marks expressive. This means all the images that the continue either considers you still should still be retained. This is this is all the images country in use plus a set up a list of images that the operator might want to manually keep in practice. We see that we have a very well so in practice.

C

One one second I see is people create a common base, images that pretty much being shared on every single machine, but then they keep fanning out of there with our further image stars. So we want to making sure that we don't believe these common images on the machines at least gave up sir operator API to do that.

A

So, just to give you guys a little context here, so we in the GC from the operator API for the very basic one. We don't any. We don't need any parameter. It means like. If you trigger the GC, we will delete all the unused image on this agent so and then depends on that. On top of that, we should have introduced a new parameter for the operator API. It is a vector of image.

A

Allow use, allow the operator to specify a couple image name, so those are the image they might be used or not currently by this agent, but we want to always keep it on this agent. So we call it an exclude images which we want to keep, and then we will do the calculation to combine those images together, and then this is the active image right here. We have vision, which means like the image we always want to keep should not be deleted.

A

B

So on so so sweetheart do you have a concrete you? So you guys sounds like you guys do have a concrete use case for that exclusion list. So.

C

The way the way the way uber said our images app is we create a pretty like slab to ten different type of base in it is roughly each one model to a different programming language, then, besides, that each of the individual micro-service would extend the base image and then put our own logic of boundaries inside this means our mean. We always have a set of non paste images we expect to so we say we excited to be keep we used by all the container monkey logic. So this means like will our our clean ups.

C

Our clean up script would probably say into the it didn't already have a cache of these five to ten images. Never carry them will be quite expensive if you believe it will be, then it's very likely they'll, be we put in a very short amount of time in future. I think.

B

Okay, so so the way you specify those exclusive images is the whole part about what they imagine. Yes,.

C

Let me slow down it's here. This is the new operator API. This can, if you leave this empty, it means default default or empty, but.

B

D

Hi, so in this case, if you upgrade the base, images doesn't mean that you have to have the operator manually delete all of them on each one of those.

A

So they need do me. Do me upgrade the image with the stand, if the same image name in tech on your registry.

D

Yeah a different different high conversion. Yes,.

A

So because if it's a potable message so with a different tag, they will be regarded as a separate image. In this case.

D

Right but the old image will it be left on the box. No.

C

For example, if you are basically, you operate a base image from version, one to version two, and if you, if you exclude images, you only put version to go for whatever layers of the version, one will can be pruned now as long as they're not being used for something else. So eventually they will go away. Well, nobody uses them all refers to them.

D

Okay, I'm still looking at.

C

So I think this is pretty much covers everything we plan to doing that. Mickey.

B

Yes, so I was more thinking about, like from our craters perspective, like how to specify J some of the images like some of the fields inside that image, probably kind of weird I cashed. What does that mean.

C

B

It will be ignored, I know it will be ignored, but yeah I guess that's probably fine. I think.

C

I think using the image type. We think that if we ever implement this to accion OCI, we don't need to go back and change this right, the stream. We need to also tell it which.

B

Uni-Y, a stream private doesn't make sense because, like a strange, prodigious photographer containers for other containers, they produce like a combination of bunch of labels, whatever I think.

C

Probably clarify in the comment here say only name is really useful. Other things yes,.

B

Have you guys took a look at the the darker prune darker image, prune command.

C

We we feature here no.

B

I'm talking about API like what do they provide I remember they provide some sort of filtering stuff, but we can do that later. I guess I just want to make sure like, like we have a dislike in API change, yeah yeah I cannot yeah, actually I. Think it's for it's! Okay! For now, what type of API you you, you you, the viewing operator aprx. Is this like an API so in the future cannot change this. If we want to move to a different, we have to be backwards compatible once we introduce this.

A

B

A

So so I mean what type of the talker prune comment API you you want yeah.

B

I'm, just looking at that I'm just thinking about, if that makes sense or no I mean darker print support, darker image print support, light a which we do support I. Can you don't specify any exclude images, this pretty much a like all images, all layers, and they have also a filter like filter, which you can do arbitrary key value pairs, like you can say, I want to delete docker images layers with step, give a label, or things like this I'm.

B

Just thinking like, if that's something that we should do in the future or not, maybe doesn't make sense, because it's very specific yeah. We should take a look to compute that, but but you guys never take a look at that. I think right now: it's okay.

C

All the mess of most metals eg I taught do too much with the labels, which seems a topic. Tucker seems like it's. Api is more heavily relying of labels of both containers, images, I, don't I, don't I, don't remember which message it can be relied on that hard labor, mostly just passing through.

D

You seeing the amount of work the operator for each role out of the basement. Could you like specify expiration time for these images, for like I, don't know one day, so it makes us can determine that all these images haven't been use in the past 24 hours. Then it could be firm if it's used by all the fastest in a real cluster.

E

That seems like I, think that feels pretty over constrained to me. It feels like the kind of thing where you should have an API, that's rich enough to be, for that to be implementable, like I was going to suggest with the prune images.

E

This is a very hardwired prune images API like it's one way to print images and there's no API to let you do it differently, and what you're talking about is another slightly different way to print images which would also be built in, but if the API gave you the eight, if you had an API that gave you information about images, these things are trivial.

C

At least a simple image list: without all the images in store in the creating pretty near future, like three to six month period like just for operators to be able to understand. What's going on in class, yeah.

B

So so, basically like a like equivalent, we talk image list like all the docker image command, that we pick a bunch of things that we think that make sense and then for operation and then probably provide API for those like lists. That's something that we definitely need. I, think yeah.

A

I think I think we can. We can do that. It should be racing. Whoo-Hoo depends on what we already catch of the image used in the metadata manager. I think we just need to expose a new API in continuances to do that. Yeah.

B

And also like, given that new CRI stuff the foundation has been committed, like I, would suggest you guys also having like CRI command to do that, like maysa yeah, it's those container image, LS things like this yeah. That would be much easier. It makes sense to me.

C

B

I think prune, probably a command to if well can be up command to write, so think. It's okay on the API for MVP, but I humanized having a CLI for that yep.

B

Given that I think Kevin I saw a bunch of CI stuff being committed, so I think now it's gonna be pretty easy to handle, see how I commence. Just writing. Just writing like several hundred laughing, not like a hundred lines of Python should be sufficient boy command.

A

So yeah so I think almost on the on the by the end of the the GCD side, I would like to mention two probably right now. Gillian and myself are trying to restore explain, is to support the dragon surrounding container, which we mentioned PST. So we don't have a solution, but I will say like a lot of user, especially existing bases user with who is using unify continuous every day.

A

My one to the ma expelled like is to be better compatible, which means they don't necessary to drink all the containers on every agent, because they're gonna be paying for if they're, already using some kind of a production, so I can see many usually they'll be expecting like we expect like we can support the GC and a solution yet, but we will keep investigate on that couple days and the second problem we have right now it is embrace we discussed yesterday. So basically right now me. This is an implementation detail.

A

So right now we we we do the mock and split, and then we did a mock me lock did not order in August or it means any new coming pulling. The crest we will post it and then we could then put off then in the queue so just waiting for the image juicy finish once they finished. We leave you clean up here on them. I, probably first resumed and continued down low in it, and then you might hit a potential race. Even if it's like right before we start approved.

A

We have some image start downloading, but they are big. They are still pulling the layers and right now we trigger the pool and we lock the store and the prune is ongoing, but the aiming down low down low, they are also ongoing. So let's say: if we have like we have like some image download. They are using the layers which is download before on the in the store, but they are unmute layers. It means we might delay those layers well by for some reasons by the pool.

A

So it is a race here and then that continuing aren't gonna fail because we are missing some layers. So right now here we are, we are, we are comparing the usability and and and the use and the operator friendly friendly choice here. It means like we can either we can either wait for all the image pool finish and all the image time low finish or the come current.

A

On the current image download finish, and then we start the image proof it means it can take up to like half an hour or longer, or we can just let the container fail. So this is two different choices. Joe and myself are discussing yesterday.

B

Hold on so what okay, first of all I, think we should not fail. The container launch, that's very unexpected, like I. Still don't understand why yeah. Maybe this is like two details that we can discuss offline. Then yeah I, don't really understand why we need to do this. Yeah.

A

This is the implementation detail. Maybe we can like first.

B

Of all I think from the API perspective or from the users perspective like failing a launch. It's not an option like we can't delay the pruning potion, often in the launch. One.

C

Problem of delaying the pruning is, if you keeps having a stream saying it's a red clay speaking. If you have a stream putting happening infinitely, you're cooler will be also delayed infinitely Annie. This is since seem similar to like, if have a read/write lock, and there is a theoretical problem that if you obtain an exclusive right lock on on something, but you keeps having other other entities of training you could the right, lock side could be starved instantly. You never know there is there.

B

Are some solution to to to make it fair right, so I'm pretty sure, there's some solution to to make sure that the right law can be acquired, no matter how many we love, you have in a pound at a time.

C

You know what one of the one of the one of information with over the way to implement this would be simply queue up any. So if what we want, when the when the first pruning enters the system, we simply to have all the image pulling to at that point. So so we make sure that, all after all, the all the thumb laws currently running down all the finishes, ruining will be the first ones in the lock and the finishes work in other.

C

B

I yeah I expect things you guys use OS removed that the pruning itself should be pretty fast. You should not block the I/o or like I, should not take too long given value in synchronously trying to remove those country.

C

All the other milliseconds.

C

B

I think I expect these, like pruning like itself to be pretty fast, given that only iOS remove OS yeah, okay yeah, so shouldn't yeah I. Think like like, like delaying those pulling sounds, sounds like a good idea.

B

Okay, so I think we have 20 minutes left. Do you have anything else on this? Okay, I have a one question when you expect this to land.

A

In our probably we already thank you, it was already lens band. It is not that maybe we might like more weeks. Okay,.

E

B

That sounds good yeah. Looking forward to this feature.

B

Sorry, what what this is done, I think we should spend some time thinking about. How do we design this on by artifact store on and then how do we do? Gcd, like I, really do reference counting or are we gonna, do mark-and-sweep, what's the protein accounts and then and have a border design.

F

On this one question I had was so the GC is always running. Is it in the sense? That's odd? Is there any operator intervention required to trigger the DC or the GC is always running.

A

So right now, in this decide dock we only have to operate the API to control the GC. But after this basic GC support lens, we might introduce an aging threat. So if we specify an age effects, so we have the heuristic project in the agent which it is the very basic one. So it means like we frequency every couple days specified by the use operator. We will trigger the GC automatically. So that's the prin which we did not have in the decider yet, but something.

F

A

F

The couple of things aren't thinking about so obviously the Opry debut is only starting on the master right. So no, that's not true. There is operating on.

C

F

Agent as well so, but then the there has to be something written to hit all the agents to actually trigger this. So that's a bit complex, but the reason I was asking that question is even if you have an agent flag. Let's say you know: the agent comes up and PC some issue with the GC I.

F

Guess the only way to start it as restart the agent by removing that particular plug and stop the GC right, let's say: there's a bug in DC where that's going and destroying images of existing containers or whatnot right, something horrible happens over there and yeah. You need to stop this. That is pretty pretty intrusive to the system. If something goes wrong with this, so the only option then is to just remove the flag and then restart agent. I guess right.

A

Yes, okay, yeah, so.

B

That that's kind of kind of reminds me another question like ID like. If AJ is not up, how do we prune the images like? That's not so you just think about like what what if the disk is already full and he cannot start agents because agency, checkpoints and metadata initially during recovery, and then you want to prune the images like is there a way to do just like do am a sociology which you do cleanup of those layers, but.

E

That's been an issue forever right, I mean this same thing, exists with Sam boxes and boxes and like agent coming up right. Listen this because we do in some cases we do a bad job of disk isolation and other agent run out of space, and you kind of manually grew up around and you could bunch of stuff not very satisfying yeah.

B

I feel like this, this kind of I mean it's at least for the image layer information. All the information is on disk. Essentially, don't need the agent to be able to process that that all we can factor out that part of the logic into a separate binary that you can just hitting that by I.

E

Think I do have I think I do follow JIRA some time ago around this saying, I wish you better do a GC in I.

C

Think operator, if I would be open in a class if I have an operating machine in a large class or in that state I will we typically do something more more brutal, simply shut up any process we can and then we move any any of these temporary. It has a dated actress like soccer star, sandbox and then try to restart your agent to see whether she'll come back.

C

I think if the diskens flow is really hard to expect other process that are still functional unless the process itself is well instrumented say it does not require any this corporation to protrude to work. Sure.

E

But I mean I guess, but the only thing along that one so I can see being reliable would be to reimage the machine from scratch like just oh this thing when I dispose nuclear from all it like as soon as you go as soon as you try to write a tool which says you know, oh, you should go and clean up. Maybe you should clean up this thing I, maybe should that thing? You really need some sort of some sort of API support for the for the file system database.

E

That missiles has so you can actually respect it in a reliable way.

B

Yeah I think this is something that would just think about next, after every PS done so.

F

This goes back at the when James was making right, I mean why not run GC as a separate process. It seems like mission critical over here and it seems independent of the agent. Why not just write it as a separate process, and then any interaction of the agent can have to happen over the process. Yeah.

B

I think we can do that. This is this, require I mean we case. We can support both right like like Adrian is running. This is gonna, be a library if agents not running will be a standalone process. Okay, something like this point. It's definitely doable just the.

C

Code into a binary speciality doable nice having that as a part of the message distribution as a helper library, yeah.

B

Yeah I think that meant to go active.

C

Running an immortal pruning, we definitely need to making sure the process interact without so we don't affect any any running containers or cookies. Oh.

B

Yeah and also I think like right now, I think Josephus was working on the standalone. Continual part of that work is making container riser itself standalone that you can launch container riser with other agents that make that very possible. We can in fact factor out a way to like just run, container Iser stand-alone and then.

A

B

That binary will be much more simplified if we have that.

C

You mean you actually can run the container. You know without metal agent. Yes,.

A

So that's ago, so we have the similar situation that, like we have the continued Iser process and not working, and then we still want to trigger the GC yeah.

B

Container Iser, as far as I know that it shouldn't checkpoint anything if no task launch it's being here, so you won't try to use. You won't try to checkpoint any information into the disk unless there's a launch from somewhere. So UF disk is full that one should be still able to boot up and then do proper cleanup, yeah.

C

So in that net standalone container launching sequence, there is something to brief to put images onto the machine before it launches yeah.

B

So you are so what do you mean sorry said: I guess because.

C

To long-term image, especially cilantro, you throw to user frog useful docker container, you typically neither image and provisioning filesystem yeah.

B

C

Is that going to be also part of that binary? Yes,.

B

So containerize it contains the image provision there right. So you will launch that, but.

C

Then we so we expect we use the same set of so all those things can share together with a long-running missiles agent. So.

B

Yeah, so container riser will be a library, so if you're running was aging, it just be a library that if it's a rainy standalone like you have a helper binary, can also initialize container riser without an agent, and that container rather can give you all information about like running containers running like all the images, if you hit proper API for that container as an interface. So that makes this very possible like see.

B

Yeah I think we can discuss that later. Offline I think it's definitely more possible right now than before and ring commenting that helped a binary on to to do like pruning image or I. Do a container list will be pretty straightforward degree.

C

That would be the system.

B

Yeah kind of yeah more than risky, it's more like container D, though, because currency only handles launching not image. Provisions in part.

C

Yeah, that's very interesting. Okay,.

B

So I think we have ten minutes. I want to leave that ten minutes to the next agenda item. So Jharna Han. Are you on the line.

B

Okay, so it sounds like he's not on the line. Yeah I think he he volunteered to do some background research on the like criminals, vm support- and this is the first checkpoint outside whether I'm going to the right direction. For the question to ask ourselves yeah, are you on the line? Jonah, Jonah hat?

B

Okay, so sounds like he's not on the line. Who can do that next next time, I think we our target to do the the VM research sync next time. That's one month check. So let's target those and also move the discussion to next time. Okay, okay sounds like we are all good anything else that guys want wants to discuss.

B

No okay thanks, everyone I think, that's it for today and I'll see you guys seem to thanks for joining thanks for presenting guitar and Gilbert. Thank you guys.