Re: [HTCondor-users] How to drain a misconfiguration node

[Greg Thain <gthain@xxxxxxxxxxx>]

On 12/17/23 01:34, David Cohen wrote:

Hi,

A misconfigured node won't accept nither drain command or reconfigure:

cm ~]# condor_drain -graceful wn245
Attempt to send DRAIN_JOBS to startd <IP_ADDRESS:9618?addrs=IP_ADDRESS-9618&alias=tech-wn245.domain&noUDP&sock=startd_4107_a38d> failed
Failed to start DRAIN_JOBS command to slot1@xxxxxxxxxxxx

wn245:~$ condor_reconfig                   
ERROR
SECMAN:2010:Received "DENIED" from server for user unauthenticated@unmapped using no authentication method, which may imply host-based security.  Our address was 'IP_ADDRESS', and server's address was 'SAME_IP_ADDRESS'.  Check your ALLOW settings and IP protocols.
Can't send Reconfig command to local master

Hi David:

The "condor_drain" command requires (condor) ADMINISTRATOR access to the node.  Assuming that you are sticking with host-based authentication, you'll want to add an appropriate ALLOW_ADMINISTRATOR line to the config file on the worker nodes.

Note that if your condor security prohibits running condor_reconfig, if you are root on the local machine, you can also kill -SIGHUP the pid of the condor_master process, and that will do the equivalent of condor_reconfig

-greg