You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
As an admin I may want to archive complete spaces and navigate to blobs by their path in the ocis blobstore. On a POSIX fs the output of ls will take longer the more blobs have been stored.
The first step to make the number of blobs more manageable is to shard them by space. The blobs should by default be stored under the same path as the space so an admin could use existing unix tools to move, backup and restore a space including metadata and blob data. The path layout should be configurable so all blobs could be put in the same directory if the spaceid was not taken into account. This is a path for a future content addressed storage, where blobs are stored by their hash. Then again, that would be a new blobstore, so it is not relevant for now.
The second step is to segment the name of blobs, similar to the nodes in the space metadata. A blob with id 582d68ea-1843-4245-9289-73e33b93044f should be stored as 58/2d/68/ea/-1/843-4245-9289-73e33b93044f. Or maybe a configurable amount of segments.
The text was updated successfully, but these errors were encountered:
As an admin I may want to archive complete spaces and navigate to blobs by their path in the ocis blobstore. On a POSIX fs the output of ls will take longer the more blobs have been stored.
The first step to make the number of blobs more manageable is to shard them by space. The blobs should by default be stored under the same path as the space so an admin could use existing unix tools to move, backup and restore a space including metadata and blob data.
The path layout should be configurable so all blobs could be put in the same directory if the spaceid was not taken into account. This is a path for a future content addressed storage, where blobs are stored by their hash.Then again, that would be a new blobstore, so it is not relevant for now.The second step is to segment the name of blobs, similar to the nodes in the space metadata. A blob with id
582d68ea-1843-4245-9289-73e33b93044f
should be stored as58/2d/68/ea/-1/843-4245-9289-73e33b93044f
. Or maybe a configurable amount of segments.The text was updated successfully, but these errors were encountered: