Graceful-exit complete on all sats but one?

[Pac]

Thank you @thepaul & @heunland & @Alexey for your fast answers and detailed investigations.
It’s nice to see you guys are still carefully looking after your software.

@thepaul

So clearly, the node is waiting until cleanup for one satellite is done before allowing the next graceful exit to complete. Which is good and fine, because it should reduce thrashing on the disk.

Well, I’d say that even though it may be good, this would be a pretty good reason to display the progress on nodes’ dashboards (or at least its status) so we’re aware when the graceful-exit is actually finished, as suggested there:

Gracefull Exit Status On the GUI

I have a node that I need to shutdown. (I have others, and will be creating a new one at a later date) I read the graceful exit guide for windows and I was able to get it started. However in the guide it states that the percent complete is legacy on the CLI and will be removed in the future. I feel that if a graceful exit is in progress it should be displayed on the node dashboard with some kind of graph showing the percent complete, and some kind of ETA clock. This way I can plan according…

Now that we know a node can take more than 30 days to exit (on its side at least?), I’m wondering if @arrogantrabbit’s comment there is still valid:

@arrogantrabbit

You can plan to power off the node 30 days after starting graceful exit. You can set a reminder in the calendar. There is no point in monitoring progress — it does not change the duration

@thepaul

I don’t see where in the code we’ve asked it to wait like that!

Could a process be stalling because a disk is too busy cleaning up some files? Most of my disks were SMR, which caused incredibly long operations sometimes (the filewalker could take more than 36 hours to complete!).