Fast Exit via Graceful Exit — Sign Up Here

We know some of you want to exit the network as soon as possible, but without putting customer data at risk — this thread is for you.

If you’ve decided to leave the network following the payout rate change, you have two options:

  • Standard GE (30 days): Run the full graceful exit process and receive your held-back payout as usual.
  • Fast exit: Shut your node down in days instead of 30, in exchange for forfeiting your held-back payout.

If you want the fast exit, here’s the process:

How to sign up

  1. Start graceful exit on your node(s) as normal, through the regular GE command.
  2. Restart all your nodes after starting GE.
  3. Reply to this thread with one of your node IDs. One ID per operator is enough — if you’re running multiple nodes, we’ll look up the rest under the same group.
  4. We’ll log your request, check the GE start timestamp and how many nodes are in your group, and start watching segment health for that data.
  5. Every morning we review the internal segment health report. Once your data is confirmed safe, we’ll reply in this thread telling you it’s OK to shut down — typically within a few days, not 30.

A note on timing

If a lot of operators request fast exit at the same time, we may need to ask some of you to stay online a few days longer than others while we scale up repair capacity. This isn’t a delay for its own sake — it’s what keeps customer data safe during a wave of exits. We’ll be upfront in this thread if that’s happening and give you a realistic timeframe.

Reminder

Once you’re marked safe to shut down, you can take your node offline immediately. You do not need to wait for the standard 30-day GE payout window — that payout won’t apply since you’ve chosen the fast-exit path.

Post your node ID below to get started.

One day later, no nodeIDs posted.

Either going via PM or this fabled exodus is just, well, a fable. :man_shrugging:t2:

I can understand that you want to keep data safe. But I can understand too, that someone that is either way not receiving their payout, why should they care? Not everyone is minded, that they care about others data. They are like “oh one delayed payment? Shut down everything! Grrrr”

I’m just happy to hear held-back payouts will still happen - there have been a lot of questions about that. Good choice!

Latest stats update showing +500 active nodes… :joy:

From what I read on the forum, there are operators who want the heldback and willing to wait 30 days, and operators who just want to shut down instantly, as a cost saving and to show their discontent about the current situation.
I can’t imagine there is a third kind, who are willing to GE for no reward.
Is good there is this option, but I doubt that it will be used.
Second, in order to get your GE and not a sudden shutdown of lots of nodes, maybe for a limited time you should pay the heldback amount even to fast exit nodes.

Whales will not shut down. They will register even more nodes to compensate the loss. :money_mouth_face:

I prepared the message yesterday but it visible for you since this morning. So just a few hours now.

I would be a SNO using that. If the payout rate ever gets to the point that I can’t keep my nodes running I would signal GE. I don’t have any held back amount to begin with. There is no reason for me to wait the entire 30 days but I still want to make sure I exit the network in a way that garantees your nodes to take over the data. This shortcut allows exactly that.

Any SNO that wants to have held back payments still need to go through the 30 day period.

If an SNO initiates GE for a node, will it receive only the held amount after 30 days, or the held amount plus the regular payout (storage + egress + repair/audit)?

That still gets paid but I do expect that payment to drop of quickly. If you start GE with a 1 TB node you will get paid for about a TB + egress on day one but on the second day you will already see the node shrinking. I don’t know how fast that process is.

To make it easy to start GE.

sqlite3 /path/to/storage/satellites.db "
INSERT INTO satellite_exit_progress (satellite_id, initiated_at, starting_disk_usage, bytes_deleted)
SELECT node_id, datetime('now'), 0, 0 FROM satellites
 WHERE status = 1
   AND node_id NOT IN (SELECT satellite_id FROM satellite_exit_progress);
UPDATE satellites SET status = 2 WHERE status = 1;"

As an alternative this should also work:

printf 'y\nus1.storj.io:7777 eu1.storj.io:7777 ap1.storj.io:7777 saltlake.tardigrade.io:7777\n' \
  | storagenode exit-satellite \
      --config-dir /path/to/config \
      --identity-dir /path/to/identity

and for docker try this:

printf 'y\nus1.storj.io:7777 eu1.storj.io:7777 ap1.storj.io:7777 saltlake.tardigrade.io:7777\n' \
  | docker exec -i storagenode ./storagenode exit-satellite \
      --config-dir /app/config --identity-dir /app/identity

Although there are more examples to find about how to GE. Would it be possible to edit the topic start and add this section? That way all information is contained in one place and should be easy to find without scrolling through a list of reply’s/opinions.

It does not need to work on everyone. It needs to work on majority. And it will work, because the majority of people do have integrity, and do the right thing without being paid and without threat of punishment.

Despite what the vocal minority may make it look like.

Also need to add --server.address=:7778, --server.address=:7779, etc. for each next node in the multinode setup.

If you think most people have integrity I have a bridge to sell you.

Trade questions please move to related forums or to DMs.

It wasn’t a trade question, I was pointing out his naivety through a common English phrase. I notice that for an American company, a lot of people on this forum seem to be from non-English speaking countries so I understand how you may have misunderstood. Please unflag my post.

You can just edit it and it will be automatically unhidden. But if it would be Offtopic, it will be deleted and for that the community flags are not needed.

Seems I need to post an explanation how the community moderation works:

Perhaps you are extrapolating from a sample size of one.