Change Details

If the local disk storage engine is configured, the webserver may try to read/write from it, and the daemons may also try to read/write from it (primarily to perform garbage collection). If the two users both have full access to the files, one of these operations may fail. Conceptually, the most consistent approach might be to write some `bin/local-files-storage-engine read/write <some path>`, and have the webserver erase as well` to that. However, that's messy and heavy and gross. The daemons could also try to make some sort of call over Conduit, conceivably. This is probably OK from a scalability point of view because the local disk engine is inherently only appropriate at small scale. However, this is unblocked by other Conduit work. Another approach might be to adjust the umask when we write stuff to the local disk engine. All of these solutions seem fairly good.

If the local disk storage engine is configured, the webserver may try to read/write from it, and the daemons may also try to read/write from it (primarily to perform garbage collection). If the two users don't both have full access to the files, one of these operations may fail. Conceptually, the most consistent approach might be to write some `bin/local-files-storage-engine read/write <some path>`, and have the webserver `sudo` to that. However, that's messy and heavy and gross. The daemons could also try to make some sort of call over Conduit, conceivably. This is probably OK from a scalability point of view because the local disk engine is inherently only appropriate at small scale. However, this is blocked by other Conduit work. Another approach might be to adjust the umask when we write stuff to the local disk engine. All of these solutions seem fairly bad.

If the local disk storage engine is configured, the webserver may try to read/write from it, and the daemons may also try to read/write from it (primarily to perform garbage collection). If the two users don't both have full access to the files, one of these operations may fail. Conceptually, the most consistent approach might be to write some `bin/local-files-storage-engine read/write <some path>`, and have the webserver erase as well`sudo` to that. However, that's messy and heavy and gross. The daemons could also try to make some sort of call over Conduit, conceivably. This is probably OK from a scalability point of view because the local disk engine is inherently only appropriate at small scale. However, this is unblocked by other Conduit work. Another approach might be to adjust the umask when we write stuff to the local disk engine. All of these solutions seem fairly goodbad.