IO Experiment
The IO experiment performs large amounts of IO requests on filesystems mounted on block devices. It can run multiple requests in parallel by spawning multiple threads (“workers”) up to the number of available cores on the target system. Depending on the parameters provided, it can perform read operations, write operations, or both.
Note
This experiment tests disk throughput, not usage. Gremlin will only use ~4KB of storage space.
Options
| Parameter |
Flag |
Default |
Version |
Description |
| Dir |
-d path |
/tmp |
1.4.0 |
The root directory for the IO experiment. |
| Workers |
-w int |
1 |
1.4.0 |
The number of IO workers to run concurrently. |
| Mode |
-m <r,w,rw> |
rw |
1.4.0 |
Do only reads, only writes, or both. |
| Block Size |
-s int |
4 |
1.4.0 |
Number of Kilobytes (KB) that are read/written at a time. |
| Block Count |
-c int |
1 |
1.4.0 |
The number of blocks read/written by workers. |
| Length |
-l int |
60 |
1.4.0 |
The length of the experiment (seconds). |
Note on cloud environments
Certain cloud providers enforce I/O limits through quotas. The throughput of this experiment may vary depending on the quotas set on the target.
Privileges required
| Privilege |
Description |
| FAULT_IO | Allows performing I/O experiments |
