Re: [HTCondor-users] Interactive submission broken on 24.0.2

[Vikrant Aggarwal <ervikrant06@xxxxxxxxx>]

Example interactive submission

12/10/24 10:48:38 (fd:5) (pid:3110052) (D_ALWAYS:2) SharedPortClient: sent connection request to local schedd for shared port id schedd_3108555_bffd
12/10/24 10:48:38 (fd:5) (pid:3110052) (D_ALWAYS:2) Response for GET_JOB_CONNECT_INFO:
ErrorString = "Failed to read address of starter for this job"
Result = false
ServerTime = 1733845718

Failed to read address of starter for this job

condor_who on the worker node shows that job is running.

condor_ssh_to_job <JOBID> works as expected. If exit, the job is getting terminated which is also expected behavior, Not sure why during interactive submission jobs are not able to take sessions to the worker node.Â

Thanks & Regards,

Vikrant Aggarwal

On Tue, Dec 10, 2024 at 10:39âAM Vikrant Aggarwal <ervikrant06@xxxxxxxxx> wrote:

Looks like that interactive job runs on slots but is not able to take the terminal session to the worker node.Â

Tried to switch the Rocky9 node to cgroup v1. No luck in behavior improvement.

Thanks & Regards,

Vikrant Aggarwal

On Mon, Dec 9, 2024 at 1:45âPM Vikrant Aggarwal <ervikrant06@xxxxxxxxx> wrote:

Strangely, for us this issue is reproducible on Rocky 8 as well where cgroup v1 are used.

Initial report is from Rocky 9 which was using cgroup v2.

Thanks & Regards,

Vikrant Aggarwal

On Mon, Dec 9, 2024 at 1:13âPM Vikrant Aggarwal <ervikrant06@xxxxxxxxx> wrote:

HelloÂExperts,

Two issues:Â

First: Why is the security file missing from 23.0.17 rpm but it's present in 24.0.1 and 24.0.2 rpm(s).

# rpm -ql condor-24.0.2-1.el9.x86_64.rpm Â| grep security
/etc/condor/config.d/00-security

# rpm -ql condor-24.0.1-1.el9.x86_64.rpm Â| grep security
/etc/condor/config.d/00-security

# rpm -ql condor-23.0.17-0.763496.el9.x86_64.rpm Â| grep security
/usr/share/doc/condor/examples/50-security

Second: interactive submission on condor worker nodes with 24.0.1 and 24.0.2

Found this bugÂhttps://opensciencegrid.atlassian.net/browse/HTCONDOR-2438

Is this patch not available in 24.0.2?

Very similar to issue mentioned inÂhttps://www-auth.cs.wisc.edu/lists/htcondor-users/2024-October/msg00060.shtml

Thanks & Regards,

Vikrant Aggarwal