Deploy HipChat Data Center manually

[ramil akhmadeev]

Hello,

We trying to deply Hipchat from OVA image, I've tryed to deploy it using following lynks:

https://confluence.atlassian.com/hc/deploying-hipchat-server-609944387.html

When I doing all that described in lynk below, on step:

Wait at least five minutes for the server to start for the first time.
Your HipChat Server's IP address (https://ip-address) will be accessible and will show you the Configure HipChat Server UI when the system is ready.

But insted the Configure HipChat Server UI, I see login screen, but I've never know the password to use, what I must to do?

Alternative I tryed to use link below, that also describes configuration from OVA image using some tools:

https://confluence.atlassian.com/hipchatdc3/deploy-hipchat-data-center-manually-909770912.html

I've configured postgres authentication on OVA image and made a configuration.json:

{
"public-dns": "dns_name_of_hip_chat_server",
"license": "*********",
"shared-storage-location": "/opt/atlassian/shared-storage-location",
"redis-server": "127.0.0.1:6379",
"database-server": "<hostip>:5432",
"database-name": "hipchat_postgres",
"database-user": "hipchat",
"database-password": "*****"
}

But this give error on hipchat datacenter restart:

admin@hipchat:~$ hipchat datacenter restart
Updating system configuration, this may take a few minutes...
Traceback (most recent call last):
File "/opt/atlassian/hipchat/sbin/_datacenter.py", line 312, in <module>
args.func(args)
File "/opt/atlassian/hipchat/sbin/_datacenter.py", line 288, in restart
if not restarter.apply_configs_and_restart():
File "/opt/atlassian/hipchat/sbin/_datacenter.py", line 177, in apply_configs_and_restart
if chef_solo.run():
File "/opt/atlassian/hipchat/sbin/_hcutils.py", line 113, in run
subprocess.check_output(cmd)
File "/usr/lib/python2.7/subprocess.py", line 573, in check_output
raise CalledProcessError(retcode, cmd, output=output)
subprocess.CalledProcessError: Command '['sudo', '/opt/atlassian/hipchat/sbin/cs']' returned non-zero exit status 1

when I running sudo /opt/atlassian/hipchat/sbin/cs it gives me error:

[2017-07-11T14:02:12+00:00] INFO: Syncing chef-solo
ERROR: could not get database connection: FATAL: password authentication failed for user "hipchat"
FATAL: password authentication failed for user "hipchat"

There was an unexpected error: 'NoneType' object has no attribute 'execute'
[2017-07-11T14:02:12+00:00] ERROR: Failed to update chef configuration from the configuration DB

.........

[2017-07-11T14:02:18+00:00] INFO: Forking chef instance to converge...
[2017-07-11T14:02:22+00:00] FATAL: Chef::Exceptions::ChildConvergeError: Chef run process exited unsuccessfully (exit code 1)
[2017-07-11T14:02:22+00:00] INFO: Attempting to enable video and generate initial values
ERROR: could not get database connection: FATAL: password authentication failed for user "hipchat"
FATAL: password authentication failed for user "hipchat"

Traceback (most recent call last):
File "/opt/atlassian/hipchat/sbin/_video.py", line 284, in <module>
hcvc = HipChatVideoCommand()
File "/opt/atlassian/hipchat/sbin/_video.py", line 45, in __init__
self.config = ConfigurationClient()
File "/opt/atlassian/hipchat/sbin/_configuration_client.py", line 44, in __init__
self._load_configuration()
File "/opt/atlassian/hipchat/sbin/_configuration_client.py", line 47, in _load_configuration
db_values = self.db.get_all("configurations")
File "/opt/atlassian/hipchat/sbin/PgPgSQL.py", line 112, in get_all
self.cursor.execute(query)
AttributeError: 'NoneType' object has no attribute 'execute'
ERROR: could not get database connection: FATAL: password authentication failed for user "hipchat"
FATAL: password authentication failed for user "hipchat"

generate_cookie ERROR: 'NoneType' object has no attribute 'execute'
[2017-07-11T14:02:22+00:00] INFO: Allow custom /etc files
[2017-07-11T14:02:22+00:00] INFO: Re-enabling monitoring
* Starting daemon monitor monit [ OK ]
Recording last chef run finished at 1499781742
[2017-07-11T14:02:22+00:00] ERROR: Error: Configuration failure

Please help us to configure evaluation instance.

[Kent Baxley]

Hi Ramil,

A few things here...

1) There is no setup wizard currently available for HipChat Data Center.  When the OVA is deployed for the first time, it is expected to see a login screen.

2) It seems you are trying to use the local postgres and redis caches on the Virtual Machine image in an attempt to make this work.  This is not supported and will not work.  As soon as you invoke the 'hipchat datacenter restart' the postgres and redis caches included on the image will be disabled.   You must configure separate hosts for a load balancer, the hipchat data center node, and a host or hosts for NFS, Postgres, and Redis.   In the near future, the Redis and Postgres services will be removed from the OVA image altogether. 

To upgrade hipchat data center when a new release comes out, you must do a wholesale replacement of the OVA with the new image.  This means that any local instances in use of Postgres and Redis and File stores will be wiped out.   This is why we require external data stores for Data Center. 

The small-scale deployment option listed below is the only supported configuration for a one-node deployment of HipChat Data Center:

https://confluence.atlassian.com/hipchatdc3/deployment-options-and-sizing-guidelines-for-hipchat-data-center-930911695.html

3) If you want to use the self-contained model from older HIpChat Server releases, we still offer the 2.2.4 version of the HipChat Server image.  HipChat Server 2.2.x will be supported and patched for critical security issues until May of 2019.  If you prefer the standalone option, you can download that below.  It also includes the startup wizard:

https://s3.amazonaws.com/hcs-release/dev/2.2.4/2017.04.24.053859/HipChat-Server-2.2.4-2017.04.24.053859.ova

[Stefano Bonacina]

Hi Kent

same problem here.

As per your first bullet point:

1) how an admin user is supposed to login to the UI on the web?

I didn't find any reference to login. Some stuff on hipchat administration, some other on OS layer application administration (BTW, I found OS user: admin - password: hipchat in the server documentation. Nothing in data center docs)

Thanks for your answer

Regards

stefano

[Kent Baxley]

Hello Stefano,

Data Center has a specific set of instructions that you need to follow to deploy it.  There is no default login until you configure one yourself, but, you must set up the environment according to our system requirements:

https://confluence.atlassian.com/hipchatdc3/hipchat-data-center-architecture-907350418.html

https://confluence.atlassian.com/hipchatdc3/deployment-options-and-sizing-guidelines-for-hipchat-data-center-930911695.html

https://confluence.atlassian.com/hipchatdc3/configuration-requirements-for-hipchat-data-center-deployment-907350433.html

Once the OVA is deployed, you must follow the instructions here to connect the Data Center node to the external data stores and load balancer:  

https://confluence.atlassian.com/hipchatdc3/deploy-hipchat-data-center-manually-909770912.htm

Only after configuring the data center and other components properly via command line can you create an initial default user account and log in to the Web UI.

This is not like HipChat Server where you are presented with a setup wizard.  We are working on a setup wizard for Data Center in the very near future, but, for now, you have to follow the manual steps linked above.  

[Stefano Bonacina]

Hi Kent

thanks for your reply.

My installation is a standalone system.

I went through the Deployment guide and got a new license.

In order to create a new configuration.json file for which parameters I had no information, I had a look at confif files under /hipchat/config, in particulat to site.json

So I create a configuration.json copying some parameters from site.json. The configuration.json file contains:
{
"public-dns": "192.168.3.52",
"license": "AAABRg0OD...",
"shared-storage-location": "localhost:/",
"redis-server": "localhost:6379",
"redis-password": "",
"database-server": "localhost:5432",
"database-name": "hipchat_postgres",
"database-user": "hipchat",
"database-password": "qSLdWXhaXfCGf45"
}

I run

hipchat datacenter configure --file ./configuration.json

with the following output:

Saving license
Updated name: hipchat
Updated location: hipchat
Updated public dns: 192.168.3.52
Updated redis credentials
Updated database credentials
Updated nfs shares

Then I run 

hipchat datacenter restart

and after some minute I got the following:

admin@hipchat:~$ hipchat datacenter restart
Updating system configuration, this may take a few minutes...
Cannot translate 'hipchat.example.com' to FQDN name -- Name or service not known
Traceback (most recent call last):
File "/opt/atlassian/hipchat/sbin/_datacenter.py", line 314, in <module>
args.func(args)
File "/opt/atlassian/hipchat/sbin/_datacenter.py", line 290, in restart
if not restarter.apply_configs_and_restart():
File "/opt/atlassian/hipchat/sbin/_datacenter.py", line 177, in apply_configs_and_restart
if chef_solo.run():
File "/opt/atlassian/hipchat/sbin/_hcutils.py", line 113, in run
subprocess.check_output(cmd)
File "/usr/lib/python2.7/subprocess.py", line 573, in check_output
raise CalledProcessError(retcode, cmd, output=output)
subprocess.CalledProcessError: Command '['sudo', '/opt/atlassian/hipchat/sbin/cs']' returned non-zero exit status 1

What am I doing wrong?

Thanks for your help

Regards

[Kent Baxley]

You cannot use localhost for the data stores in configuration.json.  Do not even attempt this.  They are not supported.  

Please look back through the architecture, sizing guides, and configuration requirements for data center:

https://confluence.atlassian.com/hipchatdc3/hipchat-data-center-architecture-907350418.html

https://confluence.atlassian.com/hipchatdc3/deployment-options-and-sizing-guidelines-for-hipchat-data-center-930911695.html

https://confluence.atlassian.com/hipchatdc3/configuration-requirements-for-hipchat-data-center-deployment-907350433.html

You absolutely must use an external source for Postgres, NFS, and Redis as well as an external Load Balancer / Reverse Proxy, even if you use only one server in your data center deployment. 

The reason this is so is because the localhost datastores are disabled when you run "hipchat datacenter restart".  Plus, the OVA does NOT have an NFS server built into it.  Pointing the shared storage location to localhost:/ will not work at all. 

In addition to that, those local instances of postgres and redis will be permanently removed from future OVA images, thus, you definitely will not be able to use these in the future.  The upgrade path for HipChat Data Center also requires that the OVA be replaced each time, so, if you acutally manange to get a data center running with local data stores, they will all go away upon upgrade and nothing has been gained. 

So, if you want to run data center with one node, you must still provide two additional machines at the bare minimum: one for the load balancer / reverse proxy, and one to host postgres, NFS, and redis.  

If all you want to do is run a stand-alone hipchat server and not set up the rest of the components, then we still offer version 2.2.4 of the HipChat Server OVA.  This will remain the "all-in-one" option and we will continue to support and patch the 2.2.x releases for critical security bugs until May 30, 2019. You can download the OVA here:

https://s3.amazonaws.com/hcs-release/dev/2.2.4/2017.04.24.053859/HipChat-Server-2.2.4-2017.04.24.053859.ova

[AhmadDanial]

Hi, Ramil.

I would like you to know that we will need to gather more information to understand the issue that you are running into with the HipChat Data Center deployment on your end. A support ticket has been created for you, so keep an eye for an update from our end. Thanks!