Hi all,
Maarten Litmaath, CERN wrote:
>On Sat, 3 Sep 2005, Torsten Harenberg wrote:
>
>>-bash-2.05b$ edg-lrc -v -i guidExists guid:foo --vo dzero --endpoint
>>http://grid-dzero-rls.physik.uni-wuppertal.de:8080/dzero/edg-local-replica-catalog/services/edg-local-replica-catalog
>>edg-local-replica-catalog starting..
>>Issuing command : guidExists
>>Parameters:
>>guid:foo
>>--vo
>>dzero
>>--endpoint
>>http://grid-dzero-rls.physik.uni-wuppertal.de:8080/dzero/edg-local-replica-catalog/services/edg-local-replica-catalog
>>Unexpected exception from server :
>>org.apache.commons.dbcp.DbcpException: java.sql.SQLException: Invalid
>>authorization specification: Access denied for user:
>>[log in to unmask] (Using password: YES)
>>
>>So obviously this has an effect, but still doesn't work.
>>
>>
>>
No the error concerning the authorization of the edg_lrc disappeared,
but another problem appeared:
Registration Failed (lfn =null, pfn =
sfn://grid-se.physik.uni-wuppertal.de/tmp/lklklk)
Reason : Cannot connect to LRC
http://grid-dzero-rls.physik.uni-wuppertal.de:8080/dzero/edg-local-replica-catalog/services/edg-local-replica-catalog
Could not create newLocalReplicaCatalog : Tried to invoke method public
java.lang.String
org.edg.data.rls.EdgLocalReplicaCatalogSoapBindingSkeleton.ping() throws
java.rmi.RemoteException with arguments . The arguments do not match
the signature.; nested exception is:
java.lang.IllegalArgumentException:
java.lang.ClassCastException@1c6e818
and for this I have no idea. Any help will be appreciated.
Thank all,
Eduard
>>>Are you sure nothing relevant (YP/NIS/DNS/...) was changed on that node?
>>>
>>Pretty sure..
>>
>>The only thing Eduard tried was to migrate to a new LRC server [...]
>>
>
>Indeed, and that exactly explains the problem. Here is what he wrote
>on the LFC mailing list:
>
>
>>I set passwords for the lrc_user and rmc_user but still I'm not able to
>>complete the migration.
>>
>
>I suppose those users did not have passwords before? Can you try making
>them passwordless again? (The MySQL server should only trust the RLS node.)
>
|