Hi We're seeing a somewhat linear rise in disk usa...
# community-help
j
Hi We're seeing a somewhat linear rise in disk usage the last 7 days. Is this normal or a sign that we need to up the memory? Or is there anything else, that can cause this?
k
If you run a lot of bulk deletes / updates or collection reloads, you have to run a compaction periodically: https://typesense.org/docs/27.1/api/cluster-operations.html#compacting-the-on-disk-database
This is a per-node API so you have to run on each node.
j
Hi Thanks, that makes a lot of sense
But is this relevant for the cloud hosted solution as well?
k
Yes, because the compaction interval is set to 7 days. If you want it more frequently, you can manage that via the API.
j
Is it fixed days? Or how can I see when it runs?
k
You will see the disk size reduce immediately.
j
Just ran the compaction on all three nodes because we got
OUT_OF_DISK
on writes. It doesn't seem to impact the disk usage though
So for now we are scaling the memory