jenkins charm appears to be hard-coded to the "jenkins" application name

[ryan-beisner]

The charm appears to be hard-coded to the "jenkins" application name somewhere.

I have the jenkins charm deployed as application name "osci-master." However, hooks fail and indicate it is trying to use application name "jenkins" for an unknown reason. I believe that is a bug.

Model  Controller         Cloud/Region             Version
temp1  temp1-serverstack  serverstack/serverstack  2.1.2

App          Version  Status   Scale  Charm          Store       Rev  OS      Notes
osci-master           error        1  jenkins        local         0  ubuntu  exposed
osci-task             unknown      2  jenkins-slave  local         2  ubuntu  

Unit            Workload  Agent      Machine  Public address  Ports     Message
osci-master/0*  error     idle       0        10.5.0.3                  hook failed: "storage-attached"
osci-task/0     unknown   idle       1        10.5.0.5                  
osci-task/1*    unknown   idle       2        10.5.0.9                  

Machine  State    DNS        Inst id                               Series  AZ
0        started  10.5.0.3   d4cdcf3c-4036-4a66-b648-7ac4ff8f0636  xenial  nova
1        started  10.5.0.5   fc3cdb86-57bd-47e9-ab2d-9cf596e8968b  xenial  nova
2        started  10.5.0.9   18d88386-4732-4b6c-9d47-2c1682723e6b  xenial  nova

2017-04-12 15:24:34 ERROR juju.worker.uniter agent.go:28 resolver loop error: preparing operation "run storage-attached (jenkins/0) hook": unknown storage "jenkins/0"
2017-04-12 15:24:34 ERROR juju.worker.dependency engine.go:547 "uniter" manifold worker returned unexpected error: preparing operation "run storage-attached (jenkins/0) hook": unknown storage "jenkins/0"

[ryan-beisner]

FYI this is cs:xenial/jenkins-4

[ryan-beisner]

Note that with this bug, I've not declared, and do not wish to declare, any sort of Juju storage.

[kramik1]

I am trying to learn and setup juju for a new project. I am seeing this with the jenkins 21 release. The logs indicate it is looking for jenkins/0 for the storage name. Is this going to be fixed anytime soon given it was almost 3 years ago? And if not, can the readme be revised to state that you need to specifically name the storage on container creation? It would save other people lots of hours and would maybe get more people to use juju. If I wasn't also playing with microk8s I might have bailed a couple of days ago.

I also used multipass to do a fresh install using the --storage jenkins=5G. I now get a filesystem 0 to machine 0: Invalid devices: Device validation failed "filesystem-0": Missing device type for device "filesystem-0" issue.

[mthaddon]

I've just done a test deployment on the LXD provider, and it seemed to work okay with no storage defined and an alternate application name (I simply ran juju bootstrap localhost lxd followed by juju deploy cs:jenkins somethingelse):

Model    Controller  Cloud/Region         Version  SLA          Timestamp
default  lxd         localhost/localhost  2.9.9    unsupported  14:06:20+02:00

App            Version  Status  Scale  Charm    Store       Channel  Rev  OS      Message
somethingelse           active      1  jenkins  charmstore  stable    34  ubuntu  Jenkins is running

Unit              Workload  Agent  Machine  Public address  Ports               Message
somethingelse/0*  active    idle   0        10.0.8.187      8080/tcp,48484/tcp  Jenkins is running

Machine  State    DNS         Inst id        Series  AZ  Message
0        started  10.0.8.187  juju-4d546d-0  focal       Running

Can you confirm if you're still affected by this with revno 34 and we can follow up? I can only apologise for the long delay in updating this issue up til now.