My experience with graceful exit

After a review of my problems with my storage node:
https://forum.storj.io/t/my-storage-node-numerous-issues

I decided to test graceful exit to kindly shutdown my storage node service.

1 - The first thing I noticed is that I could graceful exit from old satellites, but not the two new ones.
I did not paid too much attention to this, but it sounds strange that I cannot withdraw from them, even if my storage node was there before the satellites. It should be more reasonable for me, in this case, that the grant period should protect the SN against the new satellite rather than the opposite.

|2020-05-08T16:54:31.947Z|ERROR|gracefulexit:chore|worker failed|{"error": "node is not yet eligible for graceful exit: will be eligible after 2020-10-18 22:02:16.715119 +0000 +0000", "errorVerbose": "node is not yet eligible for graceful exit: will be eligible after 2020-10-18 22:02:16.715119 +0000 +0000\n\tstorj.io/storj/storagenode/gracefulexit.(*Worker).Run:96\n\tstorj.io/storj/storagenode/gracefulexit.(*Chore).Run.func1.1:88\n\tstorj.io/common/sync2.(*Limiter).Go.func1:43"}

2 - Monitoring the process (really slow), I noticed that my overall usage did not lower very much because the freed space was immediately used by the two new satellites. I did not like it, because I decided to leave the network so I do not want to be filled with data I cannot refuse.

3 - I decided to lower my allocation to the minimum 500 GB. This made the space usage to lower much quicker and stopped the incoming data from the new satellites.

4 - As of today the process is finished I successfully graceful-exited from 3 out of 6 satellites, only satellite.stefan-benten.de failed (unfortunately it was the owner of most of my data).

5 - During the process and also today my SN web dashboard has been reporting that my node has been disqualified from 2 satellites satellite.stefan-benten.de and us-central-1.tardigrade.io. The first one I may understand because the process failed. The second one surprises me because I do not understand the difference with the other two satellites successfully processed and not listed instead.

At the end of the process of 3 TB of data allocated I have been able to just halve data usage and switching off my node will cause 1.3 TB of storj data (1.6TB as reported by df) to just disappear from the network.

Any comment ?
Could I have done it better ?

Just a last question… How could I check graceful-exit payouts ?

Cheers,
d4lamar

1 Like

This is known issue, seems not fixed yet. If your graceful exit was successful and you have a receipt - you are fine.

At the moment only via API:

Not “even”, more like “because of”. Since you weren’t on these satellites for the minimum time, you couldn’t graceful exit from them. But I think it’s a bug nad the new satellites shouldn’t allow you to jpin them when you initiated a graceful exit.

Or the Earnings calculator (Update 2020-05-08: v9.2.0 - Now with surge calculation, total held back amount, payouts of held back amount and accurate node age in a redesigned satellite breakdown!)

2 Likes

Thank you.

I found the calculator very useful.

It looks like only one satellite finished by the end of April, so I should expect very small payouts for this month too.

April 2020 (Version: 9.2.0)
			TYPE		PRICE			DISK	   BANDWIDTH		PAYOUT
Upload			Ingress		-not paid-			   319.78 GB
Upload Repair		Ingress		-not paid-			    21.51 GB
Download		Egress		20   USD / TB			   256.27 GB	      5.13 USD
Download Repair		Egress		10   USD / TB			    56.73 GB	      0.57 USD
Download Audit		Egress		10   USD / TB			     1.49 MB	      0.00 USD
Disk Average Month	Storage		1.50 USD / TBm	     2.24 TBm			      3.36 USD
Disk Usage		Storage		-not paid-	     1.61 PBh
_______________________________________________________________________________________________________+
Total							     2.24 TBm	   654.29 GB	      9.05 USD
Total Surge (109.818%)									      9.94 USD

Payout and held amount by satellite:
SATELLITE	MONTH	JOINED	     HELD TOTAL	      EARNED	HELD%		 HELD	        PAID
us-central-1	    8	2019-09-01     1.44 USD	  0.3267 USD	  25%	   0.0817 USD	  0.2450 USD
	SURGE (160%)				  0.5227 USD	  25%	   0.1307 USD	  0.3920 USD
	HELD AMOUNT RETURNED	      -1.44 USD						  1.4378 USD
	AFTER RETURN		       0.00 USD	  0.5227 USD		   0.1307 USD	  1.8298 USD
	PAYOUT NOTES: Graceful Exit
europe-west-1	    8	2019-09-01     5.80 USD	  0.8031 USD	  25%	   0.2008 USD	  0.6023 USD
	SURGE (160%)				  1.2850 USD	  25%	   0.3212 USD	  0.9637 USD
europe-north-1	    1	2020-04-18     0.00 USD	  0.0000 USD	  75%	   0.0000 USD	  0.0000 USD
	SURGE (160%)				  0.0000 USD	  75%	   0.0000 USD	  0.0000 USD
asia-east-1	    8	2019-09-01     0.83 USD	  0.2304 USD	  25%	   0.0576 USD	  0.1728 USD
	SURGE (160%)				  0.3686 USD	  25%	   0.0922 USD	  0.2765 USD
saltlake	    3	2020-02-11     0.40 USD	  0.1214 USD	  75%	   0.0911 USD	  0.0304 USD
	SURGE (160%)				  0.1943 USD	  75%	   0.1457 USD	  0.0486 USD
stefan-benten	    8	2019-09-01   213.16 USD	  7.5728 USD	  25%	   1.8932 USD	  5.6796 USD
_____________________________________________________________________________________________________+
TOTAL				     221.62 USD	  9.0544 USD		   2.3243 USD	  6.7301 USD
	SURGE (109.818%)				  9.9434 USD		   2.5830 USD	  7.3604 USD
	HELD AMOUNT RETURNED	      -1.44 USD						  1.4378 USD
	AFTER RETURN		     220.19 USD	  9.9434 USD		   2.5830 USD	  8.7981 USD

I did not join them after I started the graceful exit, but they did not allocate anything before and they started allocating during the graceful exit process. This is for me a missing piece in my exit procedure or a bug. For me it was not important to receive payouts from the new satellites, but at least to have a way to notify that I would be out soon and don’t want to receive any data.

Curios, what minimum time?

The minimum time you need to be on that satellite to be able to start the graceful exit. It’s 6 month a the moment, I think.

Please note that apparently stefan-benten hasn’t reported back yet. So surge is still missing for that satellite.