mirror of
https://github.com/zeek/zeek.git
synced 2025-10-02 14:48:21 +00:00
92565d4739
util::safe_write() calls abort() in case of EAGAIN errors. This is easily observed when starting clusters with 32 workers or more. Add a custom write_message() function handling EAGAIN by retrying after a small sleep. It's not clear a more complicated poll() would be much better: The pipe might be ready for writing, but then our message might not actually fit in, resulting in another EAGAIN error. And even poll() would introduce blocking/sleeping code. Take some precautions against the stem and the supervisor dead-locking when both pipes are full by draining the other end on EAGAIN errors. Closes #3043
73 lines
2 KiB
Text
73 lines
2 KiB
Text
# Run a cluster with 64 Zeek processes and an insanely large cluster
|
|
# layout (which is sent over the supervisor <-> stem pipe for every
|
|
# Supervisor::create() call. This previously triggered an instant-abort()
|
|
# due to write() returning with EAGAIN when the pipe was filled.
|
|
|
|
# @TEST-PORT: BROKER_PORT
|
|
# @TEST-REQUIRES: test "${ZEEK_USE_CPP}" != "1"
|
|
# @TEST-EXEC: btest-bg-run zeek zeek -j -b %INPUT
|
|
# @TEST-EXEC: btest-bg-wait 60
|
|
# @TEST-EXEC: btest-diff zeek/bare-1/node.out
|
|
# @TEST-EXEC: btest-diff zeek/bare-32/node.out
|
|
|
|
# So the supervised node doesn't terminate right away.
|
|
redef exit_only_after_terminate=T;
|
|
|
|
global node_output_file: file;
|
|
global topic = "test-topic";
|
|
|
|
event do_destroy(name: string)
|
|
{
|
|
Supervisor::destroy(name);
|
|
|
|
# When no nodes are left, exit.
|
|
local status = Supervisor::status();
|
|
if ( |status$nodes| == 0)
|
|
terminate();
|
|
}
|
|
|
|
event zeek_init()
|
|
{
|
|
if ( Supervisor::is_supervisor() )
|
|
{
|
|
Broker::subscribe(topic);
|
|
Broker::listen("127.0.0.1", to_port(getenv("BROKER_PORT")));
|
|
|
|
local i = 0;
|
|
local name: string;
|
|
local cluster: table[string] of Supervisor::ClusterEndpoint;
|
|
while ( ++i <= 1024 )
|
|
{
|
|
name = fmt("bare-%d", i);
|
|
cluster[name] = [$host=127.0.0.1, $p=0/tcp, $role=Supervisor::WORKER];
|
|
}
|
|
|
|
i = 0;
|
|
while ( ++i <= 32 )
|
|
{
|
|
name = fmt("bare-%d", i);
|
|
local sn = Supervisor::NodeConfig($name=name, $directory=name, $bare_mode=T, $cluster=cluster);
|
|
Supervisor::create(sn);
|
|
}
|
|
}
|
|
else
|
|
{
|
|
Broker::peer("127.0.0.1", to_port(getenv("BROKER_PORT")));
|
|
node_output_file = open("node.out");
|
|
print node_output_file, "supervised node zeek_init()";
|
|
print node_output_file, |Cluster::nodes|, "cluster_nodes!";
|
|
print node_output_file, Cluster::nodes[Cluster::node];
|
|
}
|
|
}
|
|
|
|
event Broker::peer_added(endpoint: Broker::EndpointInfo, msg: string)
|
|
{
|
|
if ( Supervisor::is_supervised() )
|
|
Broker::publish(topic, do_destroy, Supervisor::node()$name);
|
|
}
|
|
|
|
event zeek_done()
|
|
{
|
|
if ( Supervisor::is_supervised() )
|
|
print node_output_file, "supervised node zeek_done()";
|
|
}
|