Re: [Condor-users] Enabling SSL encryption in Condor

[Pascal Jermini <pascal+condor@xxxxxxxxxxxxxxxx>]

Hi all,

> I will update the binaries, and test it out again. In case of problem I
> will get back to you!

I've successfully enabled the SSL crypto/auth layer, and I've been able
to make a few machines join the pool, all with encryption (at least
apparently).

I'm now trying to submit a job to this nice encrypted pool, but all I get from 
condor_submit is this message:

ERROR: Failed to create cluster

And then I get a mail saying that the schedd exited with status 4, followed by this message:

*** Last 20 line(s) of file SchedLog:
CryptoMethods = "3DES,BLOWFISH"
SessionDuration = "3600"
Enact = "YES"
AuthMethods = "SSL"
Subsystem = "SUBMIT"
ServerPid = 13519
RemoteVersion = "$CondorVersion: 6.7.20 Jun 21 2006 $"
Sid = "ditexsrv5:13512:1152892086:0"
ValidCommands = "60007,60011,1111,457,471"
7/14 17:48:07 (fd:11) (pid:13512) DC_AUTHENTICATE: setting sock->decode()
7/14 17:48:07 (fd:11) (pid:13512) DC_AUTHENTICATE: allowing an empty message for sock. 
7/14 17:48:07 (fd:11) (pid:13512) DC_AUTHENTICATE: Success.
7/14 17:48:07 (fd:11) (pid:13512) DaemonCore: Command received via TCP from host <XXX.XXX.XXX.XXX:56377> 
7/14 17:48:07 (fd:11) (pid:13512) DaemonCore: received command 1111 (QMGMT_CMD), calling handler (handle_q) 
7/14 17:48:07 (fd:11) (pid:13512) condor_read(): nfds=11 
7/14 17:48:07 (fd:11) (pid:13512) condor_read(): nfound=1 
7/14 17:48:07 (fd:11) (pid:13512) condor_read(): nfds=11 
7/14 17:48:07 (fd:11) (pid:13512) condor_read(): nfound=1 
7/14 17:48:07 (fd:11) (pid:13512) Got request #10002 
7/14 17:48:07 (fd:11) (pid:13512) ERROR "Socket is authenticated, but has no owner!!" at line 574 in file authentication.C 
*** End of file SchedLog

Any idea of what went wrong? Is this a problem with my pool, or rather a bug in the schedd code?

thank you in advance for any hint!

Pascal