Obsolete version of OS on the server

[aminlatifi]

The server we run the farming notify reward bot on runs with an obsolete OS version. It may have security vulnerabilities itself and also doesn't allow us to run updated versions of node.js. As we have to run our scripts with node v14 which is too old.

The server OS must be upgraded for its own security and won't be a blocker in the future for further development.

[aminlatifi]

@jainkrati

[mhmdksh]

@aminlatifi Can you dockerize this setup? So that we can easily migrate it?

CC: @geleeroyale once this is dockerized, we can easily set it up later

[aminlatifi]

It hosts two services, givpower-bot (production) and givfarm notify reward.

[geleeroyale]

thanks - we already moved the givpower-bot - we will migrate givfarm-notify after we hopefully fixed the current errors in prod

[aminlatifi]

I think it's better to finally to gelato service for these kind of things. Then we won't need to keep a server for these kind of stuff.

[aminlatifi]

The server we run the farming notify reward bot on runs with an obsolete OS version. It may have security vulnerabilities itself and also doesn't allow us to run updated versions of node.js. As we have to run our scripts with node v14 which is too old.

The server OS must be upgraded for its own security and won't be a blocker in the future for further development.

@geleeroyale @mhmdksh what's the final status of this? still notify reward is there?

[geleeroyale]

@aminlatifi I can not find a deployment or exact repo for givfarm-notify.

givpower-bot is already migrated as stated above

Can we take a look at this?

[aminlatifi]

@geleeroyale the repo is giv-token-contracts

[geleeroyale]

@aminlatifi I am sorry, but all I see in this repo are contracts and contract deployments for hardhat - I can't find any services

[aminlatifi]

@aminlatifi I am sorry, but all I see in this repo are contracts and contract deployments for hardhat - I can't find any services

There are scripts run every hour by cron job.

[geleeroyale]

@aminlatifi

I cloned the repo onto our services.giveth.io instance and tried for most of the day to get it to run - but whatever I do it won't work.

1. not every file change has been tracked with git
2. • all of the dependencies are outdated to a point that it all breaks
3. • does not seem to run successfully on node14 but is set to node14 (I know its the point of this post but I am wondering why it still seems to work on the old system)

---

after

• manually creating some script files in /deployments,
• updating package dependencies,
• switching to yarn,
• and many in-between experiments

I was able to run one of the cronjob scripts, but it fails with an error:

It would be awesome if you could clean up the repository to a point of reproducability @aminlatifi - I can't run it in the current state.

[aminlatifi]

@geleeroyale Thanks for your effort so far. If you have prepared a safe server, I can work on make it run on that and solve issues.
Wdyt?

[geleeroyale]

Do you have any update on this @aminlatifi - do you need support?

[geleeroyale]

@jainkrati This issue seems to be stuck because the deployment is totally outdated at this point. I was not able to correctly install and it seems like @aminlatifi also got stuck at some point.

My suggestion is to factor out givfarm-notify into its own repo and separate deployment

[aminlatifi]

@jainkrati This issue seems to be stuck because the deployment is totally outdated at this point. I was not able to correctly install and it seems like @aminlatifi also got stuck at some point.

My suggestion is to factor out givfarm-notify into its own repo and separate deployment

@geleeroyale Do we have a source document to tell us what is deployed where? If we have it, it would be super helpful since by looking at that we can better decide what must be deployed where.

[mohammadranjbarz]

@jainkrati This issue seems to be stuck because the deployment is totally outdated at this point. I was not able to correctly install and it seems like @aminlatifi also got stuck at some point.

My suggestion is to factor out givfarm-notify into its own repo and separate deployment

@geleeroyale Can you give me access to new server then I can deploy notify reward bot there

FYI @aminlatifi

[geleeroyale]

Access to server has been granted, please give an update on this @mohammadranjbarz

[geleeroyale]

@mohammadranjbarz Did you manage to deploy notify reward on the services instance?

[divine-comedian]

@laurenluz

Do you recognize what these services do? I think they are related to some GIVeconomy infra. If the services are important I think this should go on your radar and see if we can get some help to work on it.

[geleeroyale]

I just had to switch out some keys on givpower and givnotify (we know why!) and took this opportunity to put the scripts on the new instance as well as the cronjobs (I escaped them for now, but to activate we just need to delete the # and off we go).

However I am not sure how it looks properly, so it would be great if @mohammadranjbarz or @aminlatifi could have a look at services.giveth.io to see if everything is alright and get acquainted with the server. I am happy to be available for that to do it together.

[aminlatifi]

In my view the migration is complete, we must wait for the next automated notify rewards to be sure everything is okay.
The next automated notify reward will be on January 21, since the next one (on January 7) must be done manually as the previous one got delayed and bot won't do the action on time (needs the previous round of rewarding have been finished).

[MoeNick]

@divine-comedian
not @maryjaf nor@LatifatAbdullahi is aware of teh test scenarios for this issue, are you going to test it yourself ?

[divine-comedian]

I think is is a devOps issue so not aware if it needs to be tested or if so, how it can be tested.

@geleeroyale or @aminlatifi could provide clarity I think. If we need to test notify reward and check APRs are what we expect I think @laurenluz would know what those should be back on rewards distribution we set.

[geleeroyale]

According to @aminlatifi we have to wait for the next automated notify-reward to execute in order to evaluate of our new deployment is successful (success is expected).

And - yes it is a devOPS issue, you don't have to watch this, we will close it on our own.

[geleeroyale]

Hey @aminlatifi - how did it go? Do you think we can decommission the old server now?

[aminlatifi]

Hey @aminlatifi - how did it go? Do you think we can decommission the old server now?

Yes, it worked. We can turn off the old one

[divine-comedian]

@geleeroyale @aminlatifi if this issue is resolved let's close it!

[aminlatifi]

@geleeroyale, please update here if you have turned off and removed the server. Feel free to close it as well

[geleeroyale]

It is dead

Edit: copy of home folders was preserved