IdP clustering with memcached

[Vedran Bartonicek]

Hi,
I seem to have a problem with memcached.
Using shibboleth-identity-provider-3.2.1.

This IdP where I tried memcached works otherwise - I am using it together
with SP to
provide SSO for test apps.
Here's what I did (following instructions at
https://wiki.shibboleth.net/confluence/display/IDP30/StorageConfiguration#StorageConfiguration-MemcachedStorageService
)

1) Added a bean in global.xml, specifying own hosts's ip address
2) Modified storage service conf in idp.properties:
idp.consent.StorageService = shibboleth.MemcachedStorageService
idp.replayCache.StorageService = shibboleth.MemcachedStorageService
idp.artifact.StorageService = shibboleth.MemcachedStorageService

I have put all logging on DEBUG, but nothing memcached related came out in
idp logs. On Tomcat logs I see these:

2016-09-20 12:59:20.350 INFO net.spy.memcached.MemcachedConnection:
 Reconnecting {QA sa=/10.0.221.53:11211, #Rops=0, #Wops=0, #iq=0,
topRop=null, topWop=null, toWrite=0, interested=0}
2016-09-20 12:59:20.367 INFO net.spy.memcached.MemcachedConnection:
 Reconnecting due to failure to connect to {QA sa=/10.0.221.53:11211,
#Rops=0, #Wops=0, #iq=0, topRop=null, topWop=null, toWrite=0, interested=0}
java.net.ConnectException: Connection refused
        at sun.nio.ch.SocketChannelImpl.checkConnect(Native Method)
        at
sun.nio.ch.SocketChannelImpl.finishConnect(SocketChannelImpl.java:717)
        at
net.spy.memcached.MemcachedConnection.handleIO(MemcachedConnection.java:677)
        at
net.spy.memcached.MemcachedConnection.handleIO(MemcachedConnection.java:436)
        at
net.spy.memcached.MemcachedConnection.run(MemcachedConnection.java:1445)
2016-09-20 12:59:20.368 WARN net.spy.memcached.MemcachedConnection:
 Closing, and reopening {QA sa=/10.0.221.53:11211, #Rops=0, #Wops=0, #iq=0,
topRop=null, topWop=null, toWrite=0, interested=0}, attempt 8.


It appears that nobody is listening to port 11211 (e.g. by using netstat
-tulpn).
I am expecting that memcached client on local host (as I attempted to
configure it)
would be used for storage, but it seems like some memcached initialization
failed.

Any ideas how to debug further this issue?

I am aiming for IdP HA setup, and it appears that user sessions survive in
the cluster - in case
that I kill active IdP. (sessions are stored in browser, and memcached
configuration fails on all IdPs).
Not really sure what could go wrong in the cluster if memcached (or
something else)
is not used sharing server storage.

-Vedran
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://shibboleth.net/pipermail/users/attachments/20160920/a2630f4e/attachment.html>