-
We have created an static index. Which loops through each product and create an flat index of all the product data based on translation data, so we can easy query through translation data based on translation ids. But this type of index keeps growing in size. When we compact an database the size of the index is 1.5 GB but in one day it suddenly grows to > 2 GB. When nothing changes in the documents you should think that the size will be the same. Can somebody explain why this is/happens and if we can prevent this. |
Beta Was this translation helpful? Give feedback.
Replies: 1 comment 4 replies
-
It relates to how RavenDB does allocation of disk space. See: https://ayende.com/blog/193697-A/disk-space-management-inside-voron We are doing that on 1GB boundary once you hit the 1GB mark. When you compact, we drop the size to the minimum required. That has major advantages in terms of performance and disk optimizations. Note that RavenDB manage this internally and this isn't unbounded growth. It's merely pre-allocation and then RavenDB uses that directly. |
Beta Was this translation helpful? Give feedback.
Basically, this sort of change force us to reindex the documents. They changed.
Depending on what indexing engine you use, that means that we need to allocate more space for the new versions, and we'll delete the old versions later.
If you'll use Corax, the disk space is likely going to be more stable.