I'm sorry!
My expression is not correct.
This is not a problem with S3QL.
I just can't understand the flow of S3QL put big file(size > max_obj_size) to
backend.
Let me describe the flow again as below.
File is uploaded from client to S3QL.
|---------------------|
| File |
|---------------------|
S3QL separates the file to 3 blocks as below.
|---------------------|
|A |B | C |The size of the file is A + B + C.
|---------------------| A = B = max_obj_size
C = FileSize % max_obj_size
File is transferred from client to S3QL
|--------------------------------------->
S3QL creates object "D" which filled in zeros and have the same size with
object "C",
and PUT Object "D" to backend.
|-------------->
S3QL PUT Object "A" to backend.
|------------>
S3QL PUT Object "B" to backend.
|----------->
S3QL PUT Object "C" to backend.
|------->
S3QL DELETE Object "D"
|------|
|-----------------------------------------------------------------> Time
The flow of S3QL puts object to backend.
1) PUT Object: "D" to backend is started ( "A" is transferring from client to
S3QL cache region )
2) When the Object A finish transferring, PUT Object: "A" to backend is
started
3) When the Object B finish transferring, PUT Object: "B" to backend is
started
4) When the Object C finish transferring, PUT Object: "C" to backend is
started
5) When PUT Object: "C" is returned 201, DELETE Object: "D" is requested to
backend
I want to study below.
Although "D" is finally deleted from backend, why does S3QL create and put the
zero filled file "D" to backend ?
--
You received this message because you are subscribed to the Google Groups
"s3ql" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to [email protected].
For more options, visit https://groups.google.com/d/optout.
