Hello Andy,
I have the question and problem again. :-) So, I needed to do DNS round
robin for the redirector node because the node went sometimes down and
whole xrootd cluster was therefore unaccessible.
So, I made this change in configuration:
if named dataserver supervisor
olb.subscribe xrdstar.rcf.bnl.gov+ 3121
fi
if named redirector supervisor
odc.manager xrdstar.rcf.bnl.gov+ 3121
fi
And restarted the whole system. But now I am still getting messages on
supervisor's nodes like:
060509 13:23:01 14336 olb_Manager: Server rcas6102 already logged in.
060509 13:23:02 14336 olb_Manager: Server rcas6246 already logged in.
060509 13:23:03 14336 olb_Manager: Server rcas6291 already logged in.
060509 13:23:03 14336 olb_Manager: Server rcas6008 already logged in.
060509 13:23:04 14336 olb_Manager: Server rcas6221 already logged in.
060509 13:23:06 14336 olb_Manager: Server rcas6177 already logged in.
060509 13:23:06 14336 olb_Manager: Server rcas6353 already logged in.
060509 13:23:07 14336 olb_Manager: Server rcas6249 already logged in.
060509 13:23:08 14336 olb_Manager: Server rcas6355 already logged in.
060509 13:23:08 14336 olb_Manager: Server rcas6326 already logged in.
060509 13:23:09 14336 olb_Manager: Server rcas6375 already logged in.
For particular node rcas6326 which is in this list above, I see this
message:
060509 13:22:20 1579 olb_Server: Logged into rcas6120
060509 13:22:35 1579 olb_Open: Unable to connect socket to
rcas6280.rcf.bnl.gov; connection refused
060509 13:22:44 1579 olb_Server: Logged into rcas6120
060509 13:22:59 1579 olb_Open: Unable to connect socket to
rcas6280.rcf.bnl.gov; connection refused
060509 13:23:08 1579 olb_Server: Logged into rcas6120
060509 13:23:24 1579 olb_Open: Unable to connect socket to
rcas6280.rcf.bnl.gov; connection refused
060509 13:23:33 1579 olb_Server: Logged into rcas6120
060509 13:23:48 1579 olb_Open: Unable to connect socket to
rcas6280.rcf.bnl.gov; connection refused
060509 13:23:57 1579 olb_Server: Logged into rcas6120
060509 13:24:12 1579 olb_Open: Unable to connect socket to
rcas6280.rcf.bnl.gov; connection refused
060509 13:24:21 1579 olb_Server: Logged into rcas6120
It seems to me, that the node is trying to connect to supervisor's node
twice.
If I can say, all servers are trying to connect to to both redirectors
(under the DNS name with + sign) and both redirectors are saying too
many subscribers, try one of the supervisor's nodes. But the problem is
coming then, because I have double connection for subscription from the
same node on supervisor's node.
Do I have something wrong in my configuration ?
Thanks Pavel
|