Workload module for client.

Collaboration diagram for Workload module for client.:

Crate IO workload overview.

Crate supports two types of object I/O Operations:

1. Multiple threads writing/reading on/from the same object.
2. Each thread writing/reading on/from different objects.

For each of the above task I/O can be generated sequentially or parallely. For example, Two threads to share the object and write randomly:

NR_THREADS: 2
THREAD_OPS: 1
RAND_IO: 1
IOSIZE: 10M
BLOCK_SIZE: 32768

It is means, that crate performs next operations:

Spawns two threads.
Create a single object.
Each thread writes 4096 bytes Randomly on a shared object.

Crate I/O workload detailed description.

Workload has following parameters:

• WORKLOAD_TYPE: always 1 (I/O operations).
• WORKLOAD_SEED: initial value for pseudo-random generator (int or "tstamp").
• OPCODE: 0-CREATE, 1-OPEN, 2-WRITE, 3-READ, 4-DELETE.
• IOSIZE: - total size of I/O to perform per object.
• BLOCK_SIZE: For performance == parity group size.
• BLOCKS_PER_OP: - Number of blocks per Client operation.
• RAND_IO: Random (1) or sequential (0) IO?
• MAX_NR_OPS: - Max number of concurrent operations per thread.
• NR_OBJS - Each thread will create these many objects.
• NR_THREADS: - Number of threads.
• EXEC_TIME - time limit for executing (seconds or "unlimited").
• NR_ROUNDS: - How many times this workload to be executed.

Measurements

Currently, only execution time is measured during the test. It measures with m0_time* functions. Crate prints result to stdout when test is finished.

Logging

crate has own logging system, which based on fprintf(stderr...). (see crlog and see cr_log).

Execution time limits

There are two ways to limit execution time. The first, is operations time limit, which checks after operation executed (as described in HLD).