Ldap users authentication works, but they do not appear properly in some places. API server error (500) in resources.

[BlackPianoCat]

I have succeeded to set-up slurm-web, but there are some small issues. Particularly, concerning ldap I have succeeded to enable authentication and works quite fine (as I have tested it for my own user). In the configuration file I have the following lines,

[ldap]
uri=ldaps://ldap.eden.mini.pw.edu.pl
user_base=ou=people,dc=*,dc=*,dc=*,dc=*,dc=*
group_base=ou=group,dc=*,dc=*,dc=*,dc=*,dc=*
bind_dn=cn=admin,dc=*,dc=*,dc=*,dc=*,dc=*
cacert=/etc/certs/cacert.pem
starttls=no
user_class=account
user_name_attribute=uid
user_fullname_attribute=cn
user_primary_group_attribute=uidNumber
group_name_attribute=cn
group_object_classes=
        account
        posixGroup
        top
        shadowAccount
        edenuser

However, there are some small problems. For example, in jobs section, I can only see the (account) name in the table. Moreover, I do not see users on the filters. Apart from users, partitions are not visible in filters as well (however they are visible in the table). For some reason I cannot locate the command slurm-web-ldap-check, to check what is wrong (I do not know if it is properly installed).

Finally, I see the following error in the resources

API server error (500): The server encountered an internal error and was unable to complete your request. Either the server is overloaded or there is an error in the application.

Thank you for giving me the opportunity to ask this question.

P.S. I just want to inform that we have some users with special characters from Polish alphabet in their names. Maybe it causes some issues?

[rezib]

[…] For some reason I cannot locate the command slurm-web-ldap-check, to check what is wrong (I do not know if it is properly installed).

Did you install Slurm-web with official RPM or Deb packages? If yes, the command is installed in this path:

# /usr/libexec/slurm-web/slurm-web-ldap-check

What happens when you run this command?

Finally, I see the following error in the resources

API server error (500): The server encountered an internal error and was unable to complete your request. Either the server is overloaded or there is an error in the application.

Do you see this error above the table of nodes? If yes, did you define your RacksDB database to describe your cluster?

P.S. I just want to inform that we have some users with special characters from Polish alphabet in their names. Maybe it causes some issues?

No it shouldn't be an issue, as long as locales are using a character encoding set that supports unicode such as UTF-8.

[BlackPianoCat]

Good morning sir,

Thank you very much for your help. At this point I have the following updates:

• I found out that the script /usr/libexec/slurm-web/slurm-web-ldap-check works fine and returns me positive message. However, the usernames do not appear for some reason in table in jobs section. They appear in filters, but it does not have sense because they do not filter anything.

I use Debian system and I have installed slurm-web via apt.

• Yes, I have defined racksdb. For now i have succeeded to make drawings of the infrastructure and the data-center. Therefore, it is functional at some level. It appears that there is an authentication error in the logs of slurm-web-gateway. I am not sure why I have installed mung, and I have copied the key from the controller node, and I have generated a jwt key according to the instructions as well.

One week ago when I submitted the question, there were some conflicts with racksdb and I had to reinstall it. Probably, I tried to install it in two different ways. For now I resolved these conflicts, but there is still an authentication error.

lut 11 10:20:26 slurm-web slurm-web-gateway[264276]: Thread-182 (process_request_thread): [INFO] xxx.xxx.xxx - - [11/Feb/2025 10:20:26] "GET /api/agents/eden/jobs HTTP/1.1" 200 -
lut 11 10:20:26 slurm-web slurm-web-gateway[264276]: Thread-184 (process_request_thread): [WARNING] Unauthorized access without bearer token
lut 11 10:20:26 slurm-web slurm-web-gateway[264276]: Thread-184 (process_request_thread): [INFO] xxx.xxx.xxx - - [11/Feb/2025 10:20:26] "POST /api/agents/eden/racksdb/draw/infrastructure/eden.png?coordinates HTTP/1.1" 403 -
lut 11 10:20:26 slurm-web slurm-web-gateway[264276]: Thread-185 (process_request_thread): [INFO] xxx.xxx.xxx - - [11/Feb/2025 10:20:26] "GET /api/agents/eden/nodes HTTP/1.1" 200 -

In general, we have a very simple datacenter: a room with 3 rows with 4 racks each one of them.

In web-server, yes it is error above the table and now it changed to,

Server error: Request failed with status code 403

I hope that now my description of the problem is more precise.

Thank you very much for your will to help!

[rezib]

Hello @BlackPianoCat,

* I found out that the script `/usr/libexec/slurm-web/slurm-web-ldap-check` works fine and returns me positive message. However, the usernames do not appear for some reason in table in `jobs` section. They appear in filters, but it does not have sense because they do not filter anything.

This missing user field has nothing to do with LDAP as this information actually comes from Slurm.

Do you see the user names associated to jobs in slurmrestd API with this command?

# curl --silent --unix-socket /run/slurmrestd/slurmrestd.socket http://slurm/slurm/v0.0.40/jobs | jq '.jobs[] | {job_id,user_id,user_name}'

You can also check the user names defined in accounting database and reported by the API:

# curl --silent --unix-socket /run/slurmrestd/slurmrestd.socket http://slurm/slurmdb/v0.0.40/users | jq .users[].name

I use Debian system and I have installed slurm-web via apt.

* Yes, I have defined `racksdb`. For now i have succeeded to make drawings of the infrastructure and the data-center. Therefore, it is functional at some level. It appears that there is an authentication error in the logs of `slurm-web-gateway`. I am not sure why I have installed mung, and I have copied the key from the controller node, and I have generated a `jwt` key according to the instructions as well.

One week ago when I submitted the question, there were some conflicts with racksdb and I had to reinstall it. Probably, I tried to install it in two different ways. For now I resolved these conflicts, but there is still an authentication error.

lut 11 10:20:26 slurm-web slurm-web-gateway[264276]: Thread-182 (process_request_thread): [INFO] xxx.xxx.xxx - - [11/Feb/2025 10:20:26] "GET /api/agents/eden/jobs HTTP/1.1" 200 -
lut 11 10:20:26 slurm-web slurm-web-gateway[264276]: Thread-184 (process_request_thread): [WARNING] Unauthorized access without bearer token
lut 11 10:20:26 slurm-web slurm-web-gateway[264276]: Thread-184 (process_request_thread): [INFO] xxx.xxx.xxx - - [11/Feb/2025 10:20:26] "POST /api/agents/eden/racksdb/draw/infrastructure/eden.png?coordinates HTTP/1.1" 403 -
lut 11 10:20:26 slurm-web slurm-web-gateway[264276]: Thread-185 (process_request_thread): [INFO] xxx.xxx.xxx - - [11/Feb/2025 10:20:26] "GET /api/agents/eden/nodes HTTP/1.1" 200 -

[…]

You are actually facing this bug: #471

Update your Slurm-web packages to 4.1.0-2, restart the services, and it should be fine.

[BlackPianoCat]

• Problem with racksdb was resolved after updating the packages according to RacksDB Bearer token error #471.
• In case of slurm, yes, I see that there is an issue there (yet not sure how to fix it).

Your first command returns me indeed jobs without usernames,

{
  "job_id": 975735,
  "user_id": 1139,
  "user_name": ""
}
...

The other one returns me usernames as it should do.

[BlackPianoCat]

It appeared that sssd was not installed and this is why the uids were not resolved correctly to usernames. Thanks very much for help in troubleshooting!! :)