SSO plugin significantly delays 'stop' of Nuxeo server

(as of now) we have developed 3 plugins for Nuxeo (one for simple assets deployments, one for replacing the default bin-mgr, and one for SSO); what I have noticed is that when the SSO plugin is used (i.e. deployed to 'bundles'), the Nuxeo-server 'stop' - which typically takes 3-4 sec at the most on out system - takes so long, that eventually it times out with

“Stopping server…………………………………………..No answer from server, try to kill process 4144… Server forcibly stopped.“.

Does anyone have any idea what is it the plugin might be doing wrong, or else why this happens and how to resolve it ?

p.s. compared server.log of 2 runs - one with SSO plugin, the other without SSO plugin (where a run includes bringing Nuxeo up, then down, without doing anything in between) - and one puzzle I have noticed is that without the SSO-plugin, the localhost-startStop thread name (if I am not mistaken that 'column' it is the thread name) is consistently 'localhost-startStop-1' through out the file, where as in the run with the SSO-plugin - startup is 'localhost-startStop-1' and yet shutdown 'localhost-startStop-2'. to a Nuxeo newbie like me this does not make sense, but even if this a lead I am not sure how to proceed/what the cause might be.

p.s. also noticed that startup time is also much longer when SSO plugin is used.

0 votes

0 answers