[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [HTCondor-users] SECMAN:2011 Error with condor_status on Windows (condor v24.9.0)



Hi Aravind,

Seems like condor_status is not being authenticated correctly to the Collector. The collector log should contain more information about why it is hanging up on the tool. You can find the collector log path by running condor_config_val COLLECTOR_LOG. You can use the timestamps from the tool debug output to determine where to look in the log file.

-Cole Bollig

From: HTCondor-users <htcondor-users-bounces@xxxxxxxxxxx> on behalf of aravind.saligram@xxxxxx <aravind.saligram@xxxxxx>
Sent: Wednesday, May 21, 2025 5:31 AM
To: 'HTCondor-Users Mail List' <htcondor-users@xxxxxxxxxxx>
Subject: Re: [HTCondor-users] SECMAN:2011 Error with condor_status on Windows (condor v24.9.0)
 

Hi John, Cole

 

Thanks for the reply.

 

Use of condor: To run multiple MATLAB instances for simulations having independent parameter sweeps

First steps: Install and check condor is working on a single device (laptop in my case) as both Head Node and Worker.

Next Steps: I have 2-3 idle workstations available, make one of them the Head Node and distribute jobs. All of them are running Windows 10.

 

After running  condor_status -debug:D_SECURITY:

05/21/25 12:11:57 Win32 sysapi_get_network_device_info_raw()

05/21/25 12:11:57 SECMAN: command 5 QUERY_STARTD_ADS to collector at <192.168.178.25:9618> from TCP port 51923 (blocking).

05/21/25 12:11:57 condor_read(): Socket closed abnormally when trying to read 5 bytes from collector at <192.168.178.25:9618>, errno=10054

05/21/25 12:11:57 SECMAN: no classad from server, failing

Error: communication error

SECMAN:2011:Connection closed during command authorization. Probably due to an unknown command.

 

After running    condor_config_val -summary 

RELEASE_DIR = C:\condor

LOCAL_CONFIG_FILE = $(LOCAL_DIR)\condor_config.local

LOCAL_CONFIG_DIR = $(LOCAL_DIR)\config

ALLOW_ADMINISTRATOR = SYSTEM@* condor@password Administrator@* <user_name>@*

ALLOW_DAEMON = SYSTEM@*  condor@password

ALLOW_NEGOTIATOR = SYSTEM@* condor@password

ALLOW_READ = *

ALLOW_WRITE = *

SEC_CLIENT_AUTHENTICATION = OPTIONAL

SEC_DEFAULT_AUTHENTICATION = required

SEC_DEFAULT_ENCRYPTION = required

SEC_DEFAULT_INTEGRITY = required

SEC_READ_AUTHENTICATION = OPTIONAL

SECURITY_MODEL = 24.0

INSTALL_USER = <user_name>

CONDOR_HOST = $(FULL_HOSTNAME)

JAVA = C:\PROGRA~1\Java\JRE18~1.0_2\bin\java.exe

MaxJobRetirementTime = 2147483647

NEGOTIATOR_CONSIDER_PREEMPTION = False

DAEMON_LIST = MASTER COLLECTOR NEGOTIATOR STARTD

 

Looking forward to your reply. Thanks again for the previous assistance.

 

Regards

Aravind

 

From: HTCondor-users <htcondor-users-bounces@xxxxxxxxxxx> On Behalf Of John M Knoeller via HTCondor-users
Sent: 19 May 2025 19:37
To: htcondor-users@xxxxxxxxxxx
Cc: John M Knoeller <johnkn@xxxxxxxxxxx>
Subject: Re: [HTCondor-users] SECMAN:2011 Error with condor_status on Windows (condor v24.9.0)

 

It is normal for condor_config.local to be empty after installation,  the installer will put the configuration settings into c:\condor\condor_config  which condor_config.local can override.

 

Are you running condor_status on the same machine and as the same user that ran the installer?  The issue may be authentication rather than authorization. 

 

If condor_status is failing,  you can look in c:\condor\log\CollectorLog to see why it is refusing,  The server side of the condor_status query will have more information. 

 

To make sure we see the right messages, we need to run condor_status in a way that prints a timestamp for the commands.   so run this

 

   condor_status -debug:D_SECURITY

 

This will print the client side of the command, which might be interesting, but mostly we are looking for the timestamp on the command.

 

Then look in C:\condor\log\CollectorLog  for that timestamp range,  there should be messages in the collector log that indicate why it is refusing the condor_status query. 

 

The most likely reason is that condor_status cannot authenticate and therefore refuses to accept the command regardless of what ALLOW_READ is set to. 

 

It might also help if you run

 

   condor_config_val -summary 

 

and send us the output,  that will tell us what your effective security configuration is. 


From: HTCondor-users <htcondor-users-bounces@xxxxxxxxxxx> on behalf of aravind.saligram@xxxxxx <aravind.saligram@xxxxxx>
Sent: Monday, May 19, 2025 1:26 AM
To: htcondor-users@xxxxxxxxxxx <htcondor-users@xxxxxxxxxxx>
Subject: [HTCondor-users] SECMAN:2011 Error with condor_status on Windows (condor v24.9.0)

 

Hello to all experienced HT Condor users,

 

I am experiencing a persistent issue after installing HTCondor on Windows using the instructions from your official documentation (Install Windows as Administrator). I am using the 24.9.0 snapshot version from the repository.

After installation, when I run condor_status, I receive the following error:

Error: communication error

SECMAN:2011:Connection closed during command authorization. Probably due to an unknown command

 

My local config file was initially empty; after searching through the internet, I found this to be the ideal config, but it does not seem to work:

ALLOW_WRITE = *

ALLOW_READ = *

ALLOW_NEGOTIATOR = *

 

I have followed all the steps in the official guide, including accepting the default access permission settings during installation. Despite this, the error persists, and I am unable to get status information from the pool.

Could you please advise on how to resolve this issue? If you need any additional logs or configuration details, I am happy to provide them.

 

I initially plan to test condor on my laptop and then try extending on my local cluster of standalone windows CPU’s. But somehow, I am stuck at the very basic step.

 

Thank you for your assistance.