You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@directory.apache.org by "Alex Karasulu (JIRA)" <ji...@apache.org> on 2008/06/19 03:57:45 UTC

[jira] Updated: (DIRSERVER-1151) Cannot have multiple instance in-process listening on different sockets / ports

     [ https://issues.apache.org/jira/browse/DIRSERVER-1151?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Alex Karasulu updated DIRSERVER-1151:
-------------------------------------

    Assignee: Alex Karasulu

Will try to look into this to see if the problem is still present.  Curious why this does not prevent the mitosis replication tests from failing.  

> Cannot have multiple instance in-process listening on different sockets / ports
> -------------------------------------------------------------------------------
>
>                 Key: DIRSERVER-1151
>                 URL: https://issues.apache.org/jira/browse/DIRSERVER-1151
>             Project: Directory ApacheDS
>          Issue Type: Bug
>          Components: ldap
>    Affects Versions: 1.5.1
>         Environment: Windows, JRE 1.5*
>            Reporter: Lee Feigenbaum
>            Assignee: Alex Karasulu
>             Fix For: 1.5.3
>
>
> http://issues.apache.org/jira/browse/DIRSERVER-485 is a closed issue regarding allowing multiple directory service instances to co-exist in a single process. It was solved by multiplexing directory service instances based on the configuration instance (as given by the Configuration.JNDI_KEY in the environment) which in turn returns the proper instance via .getInstanceId().
> However, there appears to be no way to create two instances based on ServerStartupConfiguration's and multiplex from the socket (port) on which each listens to down to the proper environment (and therefore to the proper directory service). Instead, whichever directory service is created last "wins", and all LDAP requests that come in over TCP/IP end up going to that instance, regardless of which port they arrive on.
> I was going to try to patch this myself, but I'm not sure where the proper state is that would be the right point to multiplex from socket -> environment. Right now, the environment is tied to a (singleton) SessionRegistry, which is wiped out (via SessionRegistry.releaseSingleton()) when a new LdapProtocolProvider is instantiated. If SessionRegistry were changed to map from ??? to different instances, there would still need to be multiplexing code for handling messages that passes them onto the proper SessionRegistry. Right now, messages seem to only be multiplexed based on type.
> Anyway, sorry this is a bit disjointed; I only started digging into the Apache DS code yesterday to figure out why my two instances on different ports were trampling on each other. I hope this is helpful and resolvable :)

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.