Child pages
  • opHA 3 Troubleshooting

Versions Compared


  • This line was added.
  • This line was removed.
  • Formatting was changed.

Table of Contents

Peer cannot be discovered

From the master, we can run the discover using the cli tool with debug enabled to get further information: 

Code Block
bin ./opha-cli.exe act=discover url_base=http://poller username=xxxxxx password=xxxxxx debug=9

If there is any connectivity error, you can test the connectivity from the poller to the master using curl (This is an example with default credentials): 

Code Block
curl -k -d 'username=nmis&password=nm1888' -H 'DNT: 1' -H 'Accept: application/json' http://poller/omk/opCore/login

The response should be something like this: 

Code Block
{"message":"Authenticated as user nmis","ok":1}

If this is a login error, please review the credentials used are correct in the poller. You can check the file log/auth.log from the poller. 

Another common error is a misconfiguration in the poller server. The poller need to be able to generate the registry, a document with discovery information. For this, it needs to have defined both properties opha_url_base and opha_hostname (Or both, can be null if you are not using https). When the registry cannot be created you can see this error in opHA logs:

Code Block
[error] Capability Registration failed: Invalid capability document!

You can check if the registry is created in the following url:

Code Block

opHA cluster_id already exist!

Check the master and all other pollers nmis configuration for the cluster_id property. It can be found in nmis9/conf/Config.nmis. This one should be unique per server, so in case one of the pollers has a repeated one you can remove the property in the configuration file and nmis will generate a new one.

Please, notice that in case the server has nodes already, the nodes should be exported and imported again with localised_ids once the cluster_id was changed, as the nodes information won't have the same cluster_id attribute and they will be treated as remote nodes (They cannot be edited, or polled, as an example). 

401 Error from the poller

opHA uses user/password to access the registry data from the poller, but once the poller has been discover, it uses a token for authentication. So, we should have enabled the authentication method "token" in the poller. 

Check if in om/conf/opCommon.nmis we have the following (Being X 1, 2 or 3, not matter the order): 

Code Block
'auth_method_X' => 'token',

Also, the property auth_token_key should be set up in the poller configuration.

Configuring a poller over https

From the master, we can initiate discovery  of a peer using the url https://servername.   (using SSL/TLS).

But, we will not be able to query the poller as the poller will report it's url as http: .  To force the master to use HTTPS to the poller we must have the configuration item opha_url_base set to https otherwise, it won't work. 

This can be set in <omk_dir>/conf/opCommon.nmis in the poller:

Code Block
'opha_url_base' => 'https://servername'  

If we set the url to https://servername in the discover, the poller is going to send its registry data to the master, and the master will get the correct url_base for the peer from that information.  

If the opha_url_base is blank the master will swap the https:// URL for http://

Some data is not updated in the master

opHA has a new feature to synchronise only the data that has being added/modified since the last synchronisation. In case some data is not modified, we can perform a force synchronisation, adding some parameters to update only the required data types and nodes:

Code Block act=pull data_types=[nodes|latest_data|...] peers=[nodeNames] force=t