Mirror/zeek
1
0
Fork 0
mirror of https://github.com/zeek/zeek.git synced 2025-10-02 06:38:20 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions 1

Avoid race in the cluster/broker/publish-any btest

Browse source 
On very busy machines the hardwired scheduling of the ping batches could move
around among the arriving pongs, causing baseline deviations. We now wait for
each batch to complete before triggering the next one.
This commit is contained in:
Christian Kreibich 2025-04-17 14:58:43 -07:00
parent 69a1ad2c3d
commit 89780514fa
1 changed files with 9 additions and 4 deletions
Download patch file Download diff file Expand all files Collapse all files

13
testing/btest/cluster/broker/publish-any.zeek
View file

@ -56,8 +56,6 @@ event send_any()
local e = Cluster::make_event(ping, i, type_name(val), val);
Cluster::publish_hrw(Cluster::worker_pool, cat(i), e);
++i;
schedule 0.05sec { send_any() };
}
event pong(c: count, what: string, val: any)
@ -65,10 +63,17 @@ event pong(c: count, what: string, val: any)
++pongs;
print "got pong", pongs, "for ping", c, what, type_name(val), val;
# We send 5 pings in 3 different variations and
# get 4 one pong for each.
# The manager send 5 types of pings, in 3 different ways. The worker
# answers each ping in 4 ways, for a total of 60 expected pongs at the
# manager. Every batch of pings for one type involves 12 pongs.
if ( pongs == 60 )
Cluster::publish(Cluster::worker_topic, finish);
else if ( pongs > 0 && pongs % 12 == 0 )
{
# Wait for a batch to complete before sending the next.
event send_any();
}
}
event Cluster::node_up(name: string, id: string)