Hashstore Migration Guide

Vadim · October 2, 2025, 7:45pm

How much files left in blobs folder?

Walter1 · October 2, 2025, 8:18pm

The Memtable usage of one Node is now 2 GB RAM from 1.3 GB. Interestingly, it keeps climbing. Probably after each compaction.

Walter1 · October 2, 2025, 8:19pm

Just check the files themselves. They are often broken 0 kb files. I wouldn’t bother too much.

arrogantrabbit · October 2, 2025, 8:21pm

find /storj/3/storage/storage/blobs -type f -size 0 -delete

No.

konsou · October 2, 2025, 9:45pm

❯ du -hs /storj/3/storage/storage/blobs/
9.8M    /storj/3/storage/storage/blobs/

❯ find /storj/3/storage/storage/blobs/ -type f | wc -l
32

konsou · October 2, 2025, 9:47pm

Ok, deleted size 0 files. Let’s see how it goes.

arrogantrabbit · October 2, 2025, 10:02pm

Do you have any remaining files? It seems it’s just empty directions based on a puny size. It’s probably done migrating. You can turn of migreate_chor and/or delete empty directories to avoid metadata fetch spikes every ten minutes.

alpharabbit · October 3, 2025, 1:23pm

This is a well known bug. Stop node, delete used_space_per_prefix.db, start node again.

konsou · October 3, 2025, 5:01pm

This fixed the used space in dashboard.

Also disabled the migrate_chore (in addition to deleting the 0 size files and empty directories in blobs) and am not getting any more errors.

Thanks everyone!

snorkel · October 3, 2025, 5:30pm

Hold your horses! Not so fast! Where’s @alpharabbit 's beer?
… and ours…

thelastspark · October 3, 2025, 6:53pm

If the blobs folder is just filled with empty directories - are we okay to delete those after migration finished?

alpharabbit · October 3, 2025, 7:02pm

Delete everything in blobs is ok. But don’t delete blobs.

arrogantrabbit · October 3, 2025, 7:15pm

Yes. It’s in the wiki above.

dancekid:

Migration finished? Cleanup your blobs

delete files inside //storage/blobs/but dont delete blobs directory itself

bash

to find all non-zero files which could remain:
find /mnt/storagenode/blobs -type f -size +0
to remove all under /mnt/storagenode/blobs:
rm -rf /mnt/storagenode/blobs/*
zsh

to check:
ls -l /mnt/storagenode/blobs/**/*(.L1)
or
stat -f '%z %N' /mnt/storagenode/blobs/**/*(.L1)
to remove:
rm -rf /mnt/storagenode/blobs/*

thelastspark · October 3, 2025, 10:58pm

thanks, in my defence, reading is hard

Alexey · October 4, 2025, 6:26am

A post was merged into an existing topic: Scrolling in the forum

snorkel · October 5, 2025, 5:49pm

If you are on Synology, running nodes after sudo su in Docker, you activate the migration, you want to restart the node and you get this error:

Unable to find image '12--stop-timeout:latest' locally
docker: Error response from daemon: pull access denied for 12--stop-timeout, repository does not exist or may require 'docker login': denied: requested access to the resource is denied.
See 'docker run --help'.

just exit and sudo su again. You were in the wrong directory.
(I realised it after pulling the image 3 times )

Alexey · October 6, 2025, 5:04am

I would still recommend to stop and remove the container before applying any changes to the files on the bind filesystem, because docker may revert your changes. And it will revert with probability 90%, if you would stop the container but not remove it before changes, see

Ottetal · October 6, 2025, 8:01am

Phew, 1 month ago I started conversion of my nodes, and this morning all was finished. ~100 nodes with ~100TB active data has completed, and I’m now fully on Hash Store for all things StorJ. Very Cool

snorkel · October 6, 2025, 10:54am

You have 1TB/node?
I have 18 nodes with 110TB. You do it wrong.
Or you say you have 100 nodes with 100TB on each?

edortaprz · October 6, 2025, 11:15am

He has 100T of disk used.