Announcement: Storj to shut down europe-north-1 and us2

Announcements
1

Hello Storage Node Operators!

Important: This includes an announcement that we will be shutting down two satellites: europe-north-1, and us2. See below for details.

First of all thank you all so much for your engagement on the forum. The discussions up to now about SNO payouts and infrastructure costs have been productive and have given us insight into your perspective, which has strongly influenced the path we are following.

To continue down this path we have decided to sunset two of our non-production test satellites, us2 and europe-north1, over the coming months, and pay out all owed held amount from these satellites to node operators.

As you know, we are in the middle of testing new SNO payout pricing on our non-production satellites, but we have decided that we will soon extend the test to our production satellites (more on this in a subsequent announcement) and that we will no longer need all three non-production satellites.

In addition, there are several reasons why we have decided to shut down these two satellites. From an infrastructure perspective, having multiple test satellites doesnā€™t provide us with much incremental value. We believe that having a single test satellite on the public network will suffice for our needs, and shutting down the two additional satellites will alleviate some meaningful infrastructure and maintenance costs. Additionally, we have seen in the forum a desire expressed by some to be able to more quickly gracefully exit satellites and while this change doesnā€™t impact all satellites, we hope that it will address this desire to some extent.

At this point we just wanted to make the community aware of this planned change. We are still working through the exact decommissioning process and do not require any action from SNOs at this point. We will be in touch with a more detailed explanation of the process and will let you all know what if any actions need to be taken.

As always we look forward to your feedback and questions regarding this change.

Best
The Storj Team

26 Likes
2

Sounds like a reasonable decision to me. Thanks for letting us know in advance.

17 Likes
3

I hope this involves deleting all the data from nodes and not making SNOs delete it themselves. This is very crucial.

7 Likes
4

in my case, they can keep the storj for these sats and let the nodedata stay. its nearly nothingā€¦

5

us2 data on my node - 11.6GB, held amount $0.07
europe-north-1 data on my node - 4.31TB, held amount $34.9

I hope the data gets deleted and I do not need to figure out which directory to delete manually.

3 Likes
6

Does not seem nothing

image

7

260 GB on my node of 7.3 TB.
again: its like nothing for my case.
maybe other SNO are not so lucky, but how should i know?

8

in this case i hope the data will be deleted over a month (or so) time, and not as one block.

en1
en11600Ɨ1160 122 KB

you sure about the 4TB? how big is the node?

9

Turns out itā€™s 9.5TB :frowning: , previously I user the earnings script and it showed me 4.3TB, meaning 4.3TBm accumulated this month.

Total amount of data on the node currently is 28TB

image
image862Ɨ811 8.9 KB

10

i agree, thats not nothing. im sorry for you, thats sad :slightly_frowning_face:, but the node will be still a good one.

11

Oh, heh, Iā€™d prefer to run the delete manually. The way a node deletes a file makes multiple times the I/O necessary, wasteful.

3 Likes
12

This satellite pushed a lot of data to nodes a while ago. Older nodes will have a lot of it. Newer nodes basically nothing as they stopped pushing data to this satellite a while ago. So you only got data through repair.

@Bryanm Thanks for the heads up. I hope to hear more soon about the process of deleting data as this will impact end users (I personally likely have to do large migrations at some point soon and it would help if I could do that after the data is removed). Please keep us up to date about expectations and timelines. That said, I think this is a very reasonable move and would help cutting costs to operate these satellites.

4 Likes
13

I donā€™t know if it was done intentionally or it just happened, but the directories where the files are stored are named by some other format of the node id for some reason.

Also, when a node deletes data, it updates the databases. I do not know if those updates would be needed if the satellite is turned off, but Id rather everything was consistent.

As for IO, well, hopefully my node can handle it.

14

proper deletion of data is needed, for example i have 55TB of data from north-1
it would be good test for lazy file worker

3 Likes
15

That would depend on how they go about deleting the data. Technically they could wipe the segments from the satellite and let garbage collection take care of it. In which case you would be correct. But Iā€™m not sure they could send a bloom filters that matches no pieces. Normally about 10% of files that should be removed still match the bloom filters and will remain behind.

The other option would be to actually use uplink to delete the data. In which case the lazy file walker is not involved as data gets deleted as normal.

16

Iā€™m actually a bit happy to see this. Free up TB on my nodes is welcome at this point with the lower payouts with that data. I was getting close to needing to upgrade to more storage, and now i should be safe for awhile yet.

2 Likes
17

You could have exited those satellites.

18

Surely this will be updated with a specific date so that the SNO can run a graceful exit from these satellites just in time.

19

Why? The satellites are being shut down. No need for graceful exit.

2 Likes
20

just to be sure all the data from these test satellites is deleted and not miss the storage payout during ā€œthe coming monthsā€. Can you confirm that Storj labs will delete it?