[Bug 63010] mod_proxy_hcheck when health checks configured, esp. vs down service, consumes large memory

[bu...@apache.org]

https://bz.apache.org/bugzilla/show_bug.cgi?id=63010

--- Comment #3 from pawel.c <pa...@ingbank.pl> ---
Replicable also on 39, Windows Server 2012 

It is about health check memory leaks depends on frequency.

To reproduce...
(1) You have to have more than one high freq health check like...
ProxyHCExpr wsOK {%{REQUEST_STATUS} =~ /405|200/}
ProxyHCTemplate X1 hcexpr=wsOK hcmethod=GET hcuri=WHATEVER_WORKS_FOR_YOU_1
hcinterval=5 hcpasses=2 hcfails=1
<Proxy balancer://clusterFxAgregation>
BalancerMember http://127.0.0.1:10179 route=127.0.0.1 retry=0 hctemplate=X1
BalancerMember http://10.115.51.15:10179 status=+H route=10.115.51.15 retry=0
hctemplate=X1
BalancerMember http://10.115.51.27:10179 status=+H route=10.115.51.27 retry=0
hctemplate=X1
</Proxy>

ProxyHCTemplate X2 hcexpr=wsOK hcmethod=GET hcuri=WHATEVER_WORKS_FOR_YOU_2
hcinterval=5 hcpasses=2 hcfails=1
<Proxy balancer://clusterFxAgregation>
BalancerMember http://127.0.0.1:10178 route=127.0.0.1 retry=0 hctemplate=X2
BalancerMember http://10.115.51.15:10178 status=+H route=10.115.51.15 retry=0
hctemplate=X2
BalancerMember http://10.115.51.27:10178 status=+H route=10.115.51.27 retry=0
hctemplate=X2
</Proxy>

here those are defined as two virtual hosts in the same conf

(2) Some nodes must be down. For instance 127.0.0.1 in example above

(3) Observe memory. Leak spead depends on number of defined checks and their
freq. With configuration like above and around 20 virtual hosts it should be
like 10MB/min

-- 
You are receiving this mail because:
You are the assignee for the bug.
---------------------------------------------------------------------
To unsubscribe, e-mail: bugs-unsubscribe@httpd.apache.org
For additional commands, e-mail: bugs-help@httpd.apache.org