Issue 11

[joelreymont]

No description provided.

[leonsodhi]

I added a logging statement into main that prints out config.RefreshInterval and get:

2015-06-30 21:37:05 ERROR  remote_syslog.go:23 RefreshInterval: 30ns

Adding another logging statement into SetYAML suggests it never gets called. I think we're using a slightly different version of the YAML lib, but in case it's useful, this comment is present in the current src.

[leonsodhi]

This was the config:

files:
  - /home/leon/test.log
destination:
  host: HOST.papertrailapp.com
  port: PORT
  protocol: tls
new_file_check_interval: 30

[joelreymont]

I will expand the config test to include checking loaded values. It only ensures the CLI overrides are working correctly now.

[joelreymont]

Fixed

[leonsodhi]

Thanks!

[leonsodhi]

Heh, this is an interesting one. The configfile switch won't do anything because we've already loaded the config.

[joelreymont]

Useful later, perhaps, when we support sending the -HUP signal?

On Wed, Jul 1, 2015 at 10:42 PM Leon Sodhi notifications@github.com wrote:

In config.go
#78 (comment)
:

• if err = yaml.Unmarshal(file, &config); err != nil {

•   return fmt.Errorf("Could not parse the config file: %s", err)
  

• }
• self.Files = config.Files
• self.DestHost = config.Destination.Host
• self.DestPort = config.Destination.Port
• self.Protocol = config.Destination.Protocol
• self.Hostname = config.Hostname
• self.RefreshInterval = config.RefreshInterval
• self.ExcludeFiles = config.ExcludeFiles
• self.ExcludePatterns = config.ExcludePatterns
• return nil
  +}

+func (self *Config) override() error {

• configfile := pflag.StringP("configfile", "c", DEFAULT_CONFIG_FILE, "Path to config")

Heh, this is an interesting one. The configfile switch won't do anything
because we've already loaded the config.

—
Reply to this email directly or view it on GitHub
https://github.com/papertrail/remote_syslog2/pull/78/files#r33722036.

[leonsodhi]

This is definitely something we'll have to fix before merging. Many customers have manually specified a config file location in production environments, particularly when running multiple instances of r_s2 with different hostnames. We also use it in our example init.d script.

[joelreymont]

Oh, yeah, this resolves issue #11