acs::Cluster method broadcast (public)
<instance of acs::Cluster> broadcast args [ args... ]
Defined in packages/acs-tcl/tcl/cluster-procs.tcl
Send requests to all cluster peers.
- Parameters:
- args (required)
- Testcases:
- No testcase defined.
Source code: if {[ns_ictl epoch] > 0} { catch {::throttle do incr ::count(cluster:broadcast)} } # Small optimization for cachingmode "none": no need to # send cache flushing requests to nodes, when there is no # caching in place. # if {[ns_config "ns/parameters" cachingmode "per-node"] eq "none" && [lindex $args 0] in { acs::cache_flush_pattern acs::cache_flush_all ns_cache} } { # # If caching mode is none, it is expected that all # nodes have this parameter set. Therefore, there is no # need to communicate cache flushing commands. # return } if {[nsv_get cluster cluster_peer_nodes locations]} { # # During startup the throttle thread might not be started, # so omit these statistic values # if {[ns_ictl epoch] > 0} { foreach location $locations { catch {::throttle do incr ::count(cluster:sent)} set t0 [clock clicks -microseconds] :send $location {*}$args set ms [expr {([clock clicks -microseconds] - $t0)/1000}] catch {::throttle do incr ::agg_time(cluster:sent) $ms} } } else { foreach location $locations { :send $location {*}$args } } }XQL Not present: Generic, PostgreSQL, Oracle