Skip to content
Fetching contributors…
Cannot retrieve contributors at this time
93 lines (68 sloc) 3.84 KB
All Perlbal configuration commands: (case insensitive)
Admin commands:
- - - - - - - -
shutdown -- shuts down the server, killing all active connections
shutdown graceful -- closes listening sockets and stays alive until clients bleed off
create pool <name> -- create an empty pool
pool <name> add <ip[:port]> -- add a node to a pool
pool <name> remove <ip[:port]> -- remove a node to a pool
set [<pool>] <param> = <value> -- set a property on a pool (see pool-parameters.txt).
pool is optional when it was just created.
create service <name> -- create a new service
set [<service>] <param> = <value> -- set a property on a service (see service-parameters.txt)
service name is optional when service was just created.
enable <service> -- enable a service (start listening)
disable <service> -- disable a service (stops listening)
use <service_or_pool> -- sets the implied service or pool for future operations. note
that creating a service or pool object also "uses" it
header <service> remove <header> -- remove a header before going to backend node
header <service> insert <header>: <value> -- insert a header before going to backend node
load <plugin> -- load a plugin
unload <plugin> -- unload a plugin
plugins -- list plugins
xs -- show status of XS modules loaded
xs enable <module> -- turn on an already-loaded XS module
xs disable <module> -- turn off an already-loaded XS module
show service -- list all services
show service <service> -- show details of given service
show pool -- list all pools, nodes, and services using them
show pool <poolname> -- show members of a pool
reproxy_state -- dump state of reproxy status
server <param> = <value> -- set a server parameter, where param is one of:
aio_mode : one of "none", "linux" (Linux::AIO), or "ioaio" (IO::AIO)
This controls how disk IO is done asynchronously. Highly recommended
to use Linux::AIO or IO::AIO for webserving or reproxying files.
For purely reverse proxy or only reproxying URLs, none is fine.
aio_threads : number of child threads doing disk IO. (use between 2-50)
track_obj : developer option to track objects
pidfile : filename to write pidfile (no pidfile if not specified)
crash_backtrace : 1 or 0 indicating whether to perform a backtrace while the server is crashing.
Diagnostic commands:
- - - - - - - - - -
obj -- shows objects counts in scope
fd -- show file descriptors in use
proc -- show CPU usage, current time, pid, and total requests processed
nodes -- show BackendHTTP statistics for all backend nodes
node <ip[:port]> -- show BackendHTTP statistics for specified node
prof on -- enable profiling
prof off -- disable profiling
prof data -- dump profile data
uptime -- show time server was started, and current uptime
track -- dump out objects tracked, sorted by age
states [<service>]
leaks [....]
state changes
- - - - - -
The following things expand/interpolate in config files/commands
${ip:eth0} -- expands to the configured IP for interface "eth0". Probably
only works on Linux.
That is all.
Something went wrong with that request. Please try again.