Nodes uses a lot of RAM/SWAP direct after startup

Node Operators troubleshooting
21

Is this Synology?

22

Recommendations have not changed, we do not have a minimum required amount of RAM.
See Step 1. Understand Prerequisites - Storj Docs.
However, there are different cases, when you need to improve something, that’s true.
I’m have an old home media server (2013) with several disks, the ancient i7 (best for that time), 32GiB RAM and some Hyper-V and WSL2 VMs, it also used to run two docker desktop nodes with WSL2 engine and one storagenode service.
I didn’t move databases to SSD, they are still on HDDs. I enabled a badger cache for a biggest docker node, that’s all. But even it was not required, I just wanted to test it and get feedback for the Community.
These nodes suffered from issues posted there: When will "Uncollected Garbage" be deleted?, but didn’t have a usage discrepancy (except bugs when we didn’t update the trash usage correctly), however, scan on startup is enabled and working perfectly.
My nodes have had a “database locked” issues for TTL database during the stress test, I didn’t change anything. No manual interventions, they healed itself during the time.

1 Like
23

Problem is gone after update to 1.115.5

24

Or just a restart because of the update?

25

No. Did a restart with old version and disabled used spacewalker. No change.

26

No change in the usage?
But you said

27
  • Old version had high memory footprint with or without used spacewalker
  • New Version has low memory footprint without used spacewalker (did not test with used spacewalker enabled on startup)