Re: [HTCondor-users] Problem updating condor 9.8 to 10.2

[John M Knoeller <johnkn@xxxxxxxxxxx>]

Hi Silke,

 

Starting with 9.9 HTCondor has some new code to allow for a simpler wide area flock configuration.

I’m not sure why that code is impacting you but it looks like it is resulting in the condor_negotiator

attempting to contact the condor_schedd using NTSSPI rather than PASSWORD.

 

But your configuration only authorizes the PASSWORD identifier for negotiation (condor_pool@...), so the authorization is failing.

 

I would suggest you change your ALLOW_NEGOTIATOR line in config to permit the NTSSPI identifier as well as the PASSWORD identifier.

Your ALLOW_DAEMON config variable already does this, so I think you can just copy the last value from the ALLOW_DAEMON list

To the ALLOW_NEGOTIATOR list.

 

-tj

 

 

From: HTCondor-users <htcondor-users-bounces@xxxxxxxxxxx> On Behalf Of BIHLMAIER Silke via HTCondor-users
Sent: Friday, February 10, 2023 2:38 AM
To: htcondor-users@xxxxxxxxxxx
Cc: BIHLMAIER Silke <silke.bihlmaier@xxxxxxxxxxxxxxxxxxxx>
Subject: [HTCondor-users] Problem updating condor 9.8 to 10.2

 

 

Hi,

I have a successful  running condor cluster on version 9.8.1

We have authentication on PASSWORD, NTSSPI

 

I now wanted to update to 10.2 and I am using the same config files on master and node.  On 10.2  with the same configuration the authentication does not work.

What happens on 10.2: When my Master is also defined as submitter, I can submit jobs on the master, the jobs are distributed to the nodes and processed successfully.

But when one of my nodes is the submitter (which is my normal configuration), job are submitted but they are not distributed but stay on idle.

I see in the negotiator log  DENIED errors.

 

In Negotiator.lot I have the error

02/09/23 09:52:24 SECMAN: FAILED: Received "DENIED" from server for user AALEDSKPZO01$@lgs-net using method NTSSPI.

02/09/23 09:52:24 ERROR: SECMAN:2010:Received "DENIED" from server for user AALEDSKPZO01$@lgs-net using method NTSSPI.

02/09/23 09:52:24     Failed to send NEGOTIATE command to svc-de-zi-pps@xxxxxxxxxxx (<10.63.208.163:9618?addrs=10.63.208.163-9618&alias=AALEDSKBISI02.lgs-net.com&noUDP&sock=schedd_4720_8b3e>)

02/09/23 09:52:24 Failed to prefetch resource request lists for svc-de-zi-pps@xxxxxxxxxxx(<10.63.208.163:9618?addrs=10.63.208.163-9618&alias=AALEDSKBISI02.lgs-net.com&noUDP&sock=schedd_4720_8b3e>).

 

As mentioned, with the same config file 9.8.1  is working as expected, can submit jobs from the nodes and they are distributed successfully.

 

Meanwhile I did a lot of tests and I see that my  config files  working successfully on Version 9.8.1. already fail when using 9.9

 

So my question:

Was there any major change between 9.8 and 9.9 so that config files have to be updated. As configuration is good for 9.8 I don´t have a good idea at the moment what to change or check to get it to work.

 

I would appreciate help a lot.

 

Thanks, Silke

 

 

 

Silke Bihlmaier

Leica Geosystems Technologies GmbH

Friedrichstr. 100, 73430 Aalen, Germany
Phone: +49 7361 8895 4528

E-Mail: silke.bihlmaier@xxxxxxxxxxxxxxxxxxxx
Website: http://www.leica-geosystems.com

Leica Geosystems Technologies is part of Hexagon

 

Follow us on: