Dear Saleh,
We still have the simple fact that the OMERO server repeatedly sends requests and gets no response. Further, no other sites have reported this issue but OMERO LDAP usage is commonplace. There have in the past been buggy LDAP servers that had trouble with reusing established connections. Additionally there are various ways in which a stateful firewall in between could be improperly handling its connection table that would cause established inactive connections to drop. Basically: institutional IT problems are consistent with that restarting the OMERO server briefly solves the problem. For the failed requests it would thus be great to know what was logged on the LDAP server and even by firewalls along the way.
Despite the above, from the OMERO side there may yet be workarounds. Did the,
- Code: Select all
-Dcom.sun.jndi.ldap.read.timeout=5000
I mentioned make any difference? You may wish to try disabling connection pooling altogether: JVM options that may help include,
- Code: Select all
-Dcom.sun.jndi.ldap.connect.pool=false
-Dcom.sun.jndi.ldap.connect.pool.debug=all
-Dcom.sun.jndi.ldap.connect.pool.maxsize=1
so you may wish to experiment with some subset of those to see if any of them have a useful effect.
Note that for easy experimentation such options can be set from the command line, for instance,
- Code: Select all
$ bin/omero config set -- omero.jvmcfg.append -Dcom.sun.jndi.ldap.connect.pool=false
Good luck! Cheers,
Mark