lsload
Overview
lsload displays load information for cluster nodes. It is similar to the LSF lsload command. It queries the FSCHED scheduler for host resource usage and supports custom output formats and filters.
Options
Available Options
| Option | Description | Main Differences |
|---|---|---|
-a | Show all hosts, including Dynamic Cluster VM hosts. | Dynamic Cluster VM hosts are not supported; only standard LSF hosts are shown. |
-I | Show only specified load indices. | No difference. |
-l | Long format output with extra fields for I/O and external load indices. | it outputs dummy value 0; external load indices are not included. |
-w | Wide format output without truncation. | it outputs 0. |
-N | Show normalized CPU run queue length index. | Uses dummy cpu_factor value 2.5 in calculation. |
-E | Show effective CPU run queue length index. | No difference. |
-n | Show load information for the specified number of hosts. | No difference. |
-R | Show load information only for hosts that satisfy resource requirements. | Supports select with any of r15s,r1m,r15m,ut,pg,io,ls,it,tmp,swp,status; logical and arithmetic operators are not supported; other res_req are not supported. |
-V | Print version. | No difference. |
-h | Show command usage. | Provides more detailed help information. |
Default Output Fields and -l Output Fields
| Field | Description | Main Differences |
|---|---|---|
HOST_NAME | Standard host name, typically a two-component internet domain name. | No difference. |
status | Host status, possible values include ok, -ok, busy, lockW, lockU, unavail. | Implements ok, unavail, busy, lockU; -ok and lockW are not implemented. |
r15s | 15-second exponential average CPU run queue length. | No difference. |
r1m | 1-minute exponential average CPU run queue length. | No difference. |
r15m | 15-minute exponential average CPU run queue length. | No difference. |
ut | 1-minute exponential average CPU utilization, range 0–1. | No difference. |
pg | 1-minute exponential average memory paging rate (pages/sec). | No difference. |
ls | Current logged-in user count. | No difference. |
it | Host idle time (UNIX: keyboard idle; Windows: screen saver active time). | Outputs dummy value 0. |
tmp | Available space in /tmp. | No difference. |
swp | Available swap space. | No difference. |
mem | Available memory. | No difference. |
io | Disk I/O rate exponential average, hidden by default; shown with -l. | No difference. |
hostname Filters
| Filter | Description | Main Differences |
|---|---|---|
| Default (none) | Show all hosts. | No difference. |
| host_name ... | Show the selected hosts. | No difference. |
| cluster_name | Show hosts in the specified cluster. | Multi-cluster is not supported; only the current cluster is shown. |
Examples
Example 1: Show default load information for all hosts
lsload
Field notes: includes host name, status (ok/busy/unavail/lockU), CPU load indices (r15s/r1m, etc.), memory usage, and more.
Example 2: Limit to top 3 hosts
lsload -n 3
Shows only the top 3 hosts by load.
Example 3: Filter by resource conditions
# Show hosts with memory > 10000MB
lsload -R "select[mem>10000]"
Example 4: Use wide format
lsload -w
Adjusts field layout for a wider display.
Notes
Compatibility Differences (LSF vs FSCHED)
- Unsupported feature: the
-ooption for custom output is not supported; using it will error.
Parameter Conflicts
- CPU run queue options are mutually exclusive:
-Nand-Ecannot be used together. - Format option precedence: if
-land-ware both used, the last one takes effect.
Supported -R Syntax
-R uses an LSF-style resource string, for example:
"select[mem>10000]"