You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
Repository navigation
LDAP Result Code 49 - BindDN should have only one or two parts #252
Sorry it took me a few days to reply. Can you test this executable? It should address the issue (you are the second user reporting it and it's a real regression):
Sorry it took me a few days to reply. Can you test this executable? It should address the issue (you are the second user reporting it and it's a real regression):
@Fusion seems the bootleg in that share has been deleted, does that mean the official release with fix will be published soon? Or can you share another link to download the fix, so I can try on my setup, thanks!
@Fusion thanks for the new link! Currently I'm running glauth in a container, but definitely a novice here, trying to figure out how to get new binaries into container and persist them.
Also seems glauth/glauth container even doesn't have some basic cmd, e.g. ls, so hit the difficulty with my normal knowledge.
@Fusion I can report back that the bootleg (build: 20220130_213326Z) also worked in my setup, the previous "msg"="BindDN should have only one or two parts" issue fixed, I can login via ldap authentication.
@Fusion , just a side note ... what is the user is not in ou=users and the OU is deeply nested?
I would have to double check but, off the top of my head, I want to say that this scenario would only happen when forwarding queries to another LDAP server (as the built-in schema uses ou=users to allow "browse-ability" and, in that case, this check would not be active.
EDIT: lol, feel a little stupid now, as on branch dev there seems to be some GitHub workflow. :) Anyways, I'd really like to help this project to publish more frequently. So if there's anything I can do, please let me know.
@Fusion I managed to build my own binary off of the dev branch on my linuxarm64 machine using the docker.io/golang:bullseye image. So I'm no longer in need for a timely official release.
However, I feel the whole CI/CD setup for glauth should be tweaked a little to allow for more frequent releases. I haven't dug into it too much, but from the looks of it your current release process is not well-documented and seems to run on a personal / local machine. (I have no Golang experience)
I think it would be nice to rethink the branching approach for this repo and automate all of the build/publish steps. That way maintainers could focus on their code and reviewing/merging pull requests, while building and publishing releases is handled by a CI pipeline. If there is any interest within the glauth project I would gladly contribute. I have some experience with self-managed Jenkins and Drone CI pipeline, as well as, Travis CI (though they seem to have removed their free OSS offerings) and should be apply to apply this to other offerings like GitHub Actions or Circle CI, too.
I realize that this is kind of off-topic for this specific Issue, but I thought I'd test the waters before starting a new seperate Issue (Feature Request / Improvement Request).
By no means am I trying to bash the current workflow or the maintainers. English is not my first language, so if any of my remarks are impolite (or worse), please let me know. ❤️
Hi Lorenz, thanks for taking the time to think about this in depth.
Regarding feeling stupid, the only reference I could find in our own documentation of the move to github workflows is on that page: https://glauth.github.io/docs/running-travis-ci-locally.html and is a mere reference to the process so don't feel bad.
i wish we could release more often and with more confidence as, even going through the automated workflow and unit tests, there is still room for improvement. At this time, we rely on a two-step process:
When we think that we are ready for production, we create a "preview" release (that goes through automation workflows, of course)
If issues are reported, we fix them in unofficial builds and wait for the impacted folks to report victory. Since they have their own stuff to worry about, it may take them a while to find the time to test the builds.
Since stability is a higher priority that improvements (GLAuth 's use case seems to have organically evolved to support orgs in production), this process means that releases are both meatier and less frequent (the former due to the latter)
Activity
greenpau commentedon Jan 28, 2022
I think the issue is rooted in this code.
glauth/v2/pkg/handler/ldapopshelper.go
Lines 601 to 612 in e7195ed
The
johndoeDN iscn=johndoe,ou=superheros,ou=users,dc=glauth,dc=com.Once you strip the base dn, you are left with
cn=johndoe,ou=superheros,ou=users.glauth/sample-simple.cfg
Line 36 in e26219c
The
cn=johndoe,ou=superheros,ou=userswill have 3 parts, as opposed to two.Thus, the logic would never pass when you have many nested groups.
Fusion commentedon Jan 29, 2022
Sorry it took me a few days to reply. Can you test this executable? It should address the issue (you are the second user reporting it and it's a real regression):
https://file.io/vfSQH4zQnVs3
greenpau commentedon Jan 29, 2022
@Fusion , this one works 👍
Fix #246 and #252
joace commentedon Feb 8, 2022
@Fusion seems the bootleg in that share has been deleted, does that mean the official release with fix will be published soon? Or can you share another link to download the fix, so I can try on my setup, thanks!
Fusion commentedon Feb 8, 2022
@joace if you are curious, I am awaiting another confirmation that this version works. You should be able to download it from https://mega.nz/file/ltgVhQJT#Kc9OBPGABc2Z1XrTFcDEXi8KGOjNwD8-SvR_emzePPg
greenpau commentedon Feb 8, 2022
@Fusion , just a side note ... what is the user is not in
ou=usersand the OU is deeply nested?joace commentedon Feb 8, 2022
@Fusion thanks for the new link! Currently I'm running glauth in a container, but definitely a novice here, trying to figure out how to get new binaries into container and persist them.
Also seems glauth/glauth container even doesn't have some basic cmd, e.g. ls, so hit the difficulty with my normal knowledge.
joace commentedon Feb 12, 2022
@Fusion I can report back that the bootleg (build: 20220130_213326Z) also worked in my setup, the previous "msg"="BindDN should have only one or two parts" issue fixed, I can login via ldap authentication.
Fusion commentedon Feb 13, 2022
I would have to double check but, off the top of my head, I want to say that this scenario would only happen when forwarding queries to another LDAP server (as the built-in schema uses
ou=usersto allow "browse-ability" and, in that case, this check would not be active.lbarnkow commentedon Feb 16, 2022
@Fusion Any idea when the fix to the
BindDN should have only one or two partsis generally available in a proper release?Fusion commentedon Feb 19, 2022
I am thinking this week-end, as I got confirmation from multiple sources that the fix works well.
lbarnkow commentedon Feb 21, 2022
EDIT: lol, feel a little stupid now, as on branch
devthere seems to be some GitHub workflow. :) Anyways, I'd really like to help this project to publish more frequently. So if there's anything I can do, please let me know.@Fusion I managed to build my own binary off of the
devbranch on mylinuxarm64machine using thedocker.io/golang:bullseyeimage. So I'm no longer in need for a timely official release.However, I feel the whole CI/CD setup for glauth should be tweaked a little to allow for more frequent releases. I haven't dug into it too much, but from the looks of it your current release process is not well-documented and seems to run on a personal / local machine. (I have no Golang experience)
I think it would be nice to rethink the branching approach for this repo and automate all of the build/publish steps. That way maintainers could focus on their code and reviewing/merging pull requests, while building and publishing releases is handled by a CI pipeline. If there is any interest within the glauth project I would gladly contribute. I have some experience with self-managed Jenkins and Drone CI pipeline, as well as, Travis CI (though they seem to have removed their free OSS offerings) and should be apply to apply this to other offerings like GitHub Actions or Circle CI, too.
I realize that this is kind of off-topic for this specific Issue, but I thought I'd test the waters before starting a new seperate Issue (Feature Request / Improvement Request).
By no means am I trying to bash the current workflow or the maintainers. English is not my first language, so if any of my remarks are impolite (or worse), please let me know. ❤️
Fusion commentedon Feb 21, 2022
Hi Lorenz, thanks for taking the time to think about this in depth.
Regarding feeling stupid, the only reference I could find in our own documentation of the move to github workflows is on that page: https://glauth.github.io/docs/running-travis-ci-locally.html and is a mere reference to the process so don't feel bad.
i wish we could release more often and with more confidence as, even going through the automated workflow and unit tests, there is still room for improvement. At this time, we rely on a two-step process:
Since stability is a higher priority that improvements (GLAuth 's use case seems to have organically evolved to support orgs in production), this process means that releases are both meatier and less frequent (the former due to the latter)
Dev (#254)
joace commentedon Mar 10, 2022
@Fusion Is there any plan to update docker image with v2.1.0 release?
Fix/docker latest tag (#260)
Dev -> Master after v1/v2 cleanup (#261)
Feature/dbsshkeys (#262)
Dev -> Master get rid of old Docker files (#288)