@dancekid is working on a wiki that explains some of these options and recommendations:
2 Likes
I’m starting migrating to hashstore but I’ve a curiosity.
Will the Filewalker process still be necessary?
I’m sorry if it will be a silly question.
Hello @gnegnus,
Yes you still need it. When the migration will be completed, you may disable it.
Interesting! I finished migrating 3 nodes and there was no problem with the free space. The dashboard show the correct occupied space during and after the migration.
I believe there were some code impruvments since few months ago.
So I didn’t need to delete that db.
may be you didnt restarted node during migration this time
I’m not sure of this, but maybe.
I didn’t, but I don’t know about updates.
Is there any news about the hashstore rollout so far?
How many nodes are converted up to now?
Is there any performance gain observable?
Other isses?
What’s the way ahead?
rollout is paused now, they wait 1.139 version update, as hashstore eat RAM as hungry on windows, 1.139 reduce RAM usage.
2 Likes
how can we check how far our nodes are with migrating? and should we do something or does this happen automagically?
Hello @mrkeyboardcommando,
Welcome back!
The simplest one by measuring the blobs and hashstore folders.
The first one should decrease the second one grow.
Please note, the first stage is to enable to accept all uploads to hashstore, so you likely would just note that hashstore just grow instead of piecestore.
1 Like
yeah i see data going into the hashstore folder. but do we need to flip a switch or… does this happen eventually automatically.
It will happen automatically one day. But you can flip all switches by hand any time to speedup things.
3 Likes
Thanks, i’ll try it on an smaller node then.
Let the startup scan finish and take note of the number of pieces you have.
After starting hashstore migration, you will see in log how many pieces have migrated. Look for this:
docker logs storagenode 2>&1 | grep "used-space-filewalker"
docker logs storagenode 2>&1 | grep "piecemigrate:chore"
1 Like
smallest node is underway. well see.
11 posts were merged into an existing topic: Single node and multinode dashboards shows different values after migration to hashstore
Looks like such a feature is on the way:
1 Like