Graceful Exit Guide

deathlessdd · April 9, 2020, 6:33pm

Are you doing it from powershell or Command prompt?
try running this

`docker exec -it storagenode /app/storagenode exit-satellite --config-dir /app/config --identity-dir /app/identity`

`docker exec -it storagenode /app/storagenode exit-status --config-dir /app/config --identity-dir /app/identity`

sparhawk4 · April 9, 2020, 7:51pm

I had tried both.

Attempting this:

`docker exec -it storagenode /app/storagenode exit-satellite --config-dir “D:\Storjv3-store” --identity-dir “C:\Users\user\AppData\Roaming\Storj\Identity\storagenode”`

returned this in Command Prompt:

‘`docker’ is not recognized as an internal or external command,
operable program or batch file.

and just this in Powershell:

>>

deathlessdd · April 9, 2020, 7:58pm

You arent running it correctly, You need to run exacly as I have put you dont need to put any location.

sparhawk4 · April 9, 2020, 8:06pm

Oh. OK, putting this into powershell:

`docker exec -it storagenode /app/storagenode exit-satellite --config-dir /app/config --identity-dir /app/identity`

gets us further, but it still seems to choke on the forward slash for indentity.cert:

2020-04-09T19:59:48.190Z INFO Configuration loaded from: /app/config/config.yaml
2020-04-09T19:59:48.192Z FATAL file or directory not found error: open /app/identity`/identity.cert: no such file or directory

However, just putting in the original command works in Powershell!

docker exec -it storagenode /app/storagenode exit-satellite --config-dir /app/config --identity-dir /app/identity

I never tried that initially because there were instructions for Linux, and instructions for Windows GUI, so I incorrectly mashed those two up. Thanks again for helping me work through this!

deathlessdd · April 9, 2020, 8:07pm

Yeah its exactly the same as linux when its running docker.

deb8 · April 17, 2020, 12:15am

Hi all,
Any chance of explaining a little more what the following means???
Thanks in advance

Domain Name Node ID Percent Complete Successful Completion Receipt
satellite.stefan-benten.de:7777 118UWpMCHzs6CvSgWd9BfFVjw5K9pZbJjkfZJexMtSkmKxvvAW 0.00% N 0a47304502205070a80c8db691ac877609cf0fd6e496a5026dfe8d5da93cb85fd7e7e8ef4076022100afb21eaac9b14845892006f784936dc79e638fd47c1e4a4a5d39521e7f93d72e1a20004ae89e970e703df42ba4ab1416a3b30b7e1d8e14aa0e558f7ee2680000000022206bbfde33c974001aa97666637c4f3ebaa6db3e5f88e75c8544b215e7000000002a0b088de3d1f40510a8abfc2d

deathlessdd · April 17, 2020, 12:19am

If it completed it would say 100% and say Y next to it. Is your GE still running you shouldnt have any receipt next to it less it completely failed GE

deb8 · April 17, 2020, 12:23am

so, what can I do? any procedures to follow?
GE is still running for the other satellites…

Domain Name Node ID Percent Complete Successful Completion Receipt
satellite.stefan-benten.de:7777 118UWpMCHzs6CvSgWd9BfFVjw5K9pZbJjkfZJexMtSkmKxvvAW 0.00% N 0a47304502205070a80c8db691ac877609cf0fd6e496a5026dfe8d5da93cb85fd7e7e8ef4076022100afb21eaac9b14845892006f784936dc79e638fd47c1e4a4a5d39521e7f93d72e1a20004ae89e970e703df42ba4ab1416a3b30b7e1d8e14aa0e558f7ee2680000000022206bbfde33c974001aa97666637c4f3ebaa6db3e5f88e75c8544b215e7000000002a0b088de3d1f40510a8abfc2d
asia-east-1.tardigrade.io:7777 121RTSDpyNZVcEU84Ticf2L1ntiuUimbWgfATz21tuvgk3vzoA6 44.05% N N/A
us-central-1.tardigrade.io:7777 12EayRS2V1kEsWESU9QMRseFhdxYxKicsiFmxrsLZHeLUtdps3S 40.83% N N/A
europe-west-1.tardigrade.io:7777 12L9ZFwhzVpuEKMUNUqkaTLGzwY9G24tbiigLiXpmZWKwmcNDDs 3.94% N N/A
saltlake.tardigrade.io:7777 1wFTAgs9DP5RSnCqKV1eLf6N9wtk4EAtmN5DpSxcs8EjT69tGE 0.00% N N/A

deathlessdd · April 17, 2020, 12:31am

Well you cannot stop it from running you have to let it finish. But it looks like you failed GE for 118 satellite tho.

cml · April 18, 2020, 12:54am

I have completed gracefull exit for 4 satellites in one of the nodes I have but saltlake is still remaining. I’m interested in recover the space this particular node is taking.

If I delete the node and its data will I receive the payout and hold amount for the other 4 satellites that I have successfully exited ? (I have the completion receipt for these)

EDIT: Just to be clear, I know it isn’t good to delete a node and its data like this. But given that saltlake is used for network testing I think it might be OK to just delete the data. No real clients will be impacted by this.

deathlessdd · April 18, 2020, 1:23am

I would say just delete everything if your just trying to get rid of the node you wont be able to GE on salt lake since its a new satellite tho.

cml · April 18, 2020, 1:26am

I want to get rid of the node so I can get some space in the disk. I don’t mind be disqualified on salt lake, just want to be sure I will get the hold amount for the other satellites if I do it.

deathlessdd · April 18, 2020, 1:35am

Yeah you will each satellite holds there own amounts.

cml · April 18, 2020, 1:37am

Thanks on the clarification @deathlessdd!

lbndev · April 18, 2020, 5:26am

This is a legitimate situation for which I think the 6/15 months minimum for performing a graceful exit is against the network’s interest.

When a honest SNO must permanently stop his node after 3 months (or 13 ?), for whatever legitimate reason (predicted hardware failure, need more disk space, 6 month vacation, …), he can’t GE. He would be willing to flush his data back into the network for free, but he isn’t allowed to do so. His only option is to get DQed and lose the escrow. So what will he do ? Just switch the node down, delete the data, and let the network bother with the repairs. Which will be costly. Actually the repairs will probably cost more to the network, than the actual escrow amount “forfeited” by the SNO.

I understand this rule was made to punish a cheater who may want to GE every 3 month to get the (low) escrow he may have accumulated in these months. But what next ? He would start a new node, wait a month to be vetted, and probably one more month before the node is full and starts to have the egress trafic for only a few $. IMO there’s no point in cheating this way, as simply keeping the node online would provide more income.

Here’s my suggestion:

allow GE at any time. No minimum.
when GE is successfully completed, pay the SNO x% of his escrow, where x depends on how much time the node has been on the network : 0% for < 3 months, 25% for < 6 months, 50% for < 9 months, 75% for < 15 months, and 100% above. These numbers look familiar on purpose.

This way “goodwill” SNOs are able to exit the network whenever they need to without causing repair costs to the network, and “evil” SNOs still make no profit in trying to game the system.

nerdatwork · April 18, 2020, 7:06am

You should make a post on ideas.storj.io describing the details or copy-paste this post so devs can have a look at it.

naxbc · May 12, 2020, 5:21pm

I´m going to do a GE on one of my SNO´s. The internet connection just sucks, and before I get disqualified, I´m starting over when my mother has a decent connection.
So my question is:
I should “docker stop -t 300 storagenode” before “storagenode exit-satellite” right?

BrightSilence · May 12, 2020, 5:26pm

I haven’t run it myself, but I’m almost certain you should not stop the node before running it. Reason being that your node is still running. It’s even still serving downloads for data still on your node during the graceful exit. It’s just an additional process your node would be working on.

naxbc · May 12, 2020, 5:28pm

Ok, thank you. Will check on it and just still leave it running.

SGC · May 12, 2020, 6:19pm

doing a GE is pretty difficult as far i i’ve understood it…
looked through the requirements of it… i would hate to go through that…
but i’m sure a lot of people need to fail before it gets changed… good luck