On Wed, Mar 26, 2008 at 04:15:10PM +0100, Julius Volz wrote:
> Hi,
> 
> This patch (already reviewed by Horms) implements a continuous logging
> mechanism for real service checks in ldirectord. This is useful for
> analyzing when and why real services were down or for statistical purposes.
> 
> A log file may be specified for each virtual section using a new
> "monitorfile = <filename>" directive. Default is no logging. The log format
> is as follows:
> 
> <unix timestamp>|<pid>|<real_service_id>|<status>|<received / message>
> 
> The "<received / message>" part is service-type specific and may contain the
> contents received from the service or an error message. Currently, only
> check_http() sets this field to something useful, default is "-". Adding
> useful messages for other service checks is encouraged ;)
> 
> An example of a log line:
> 1206443690|16083|negotiate:http:tcp:172.xx.xx.xx
> :8001:::1:gate:\/ok\.html:OK|up|OK

Thanks, this change looks good to me. I will merge it.

-- 
Horms

_______________________________________________________
Linux-HA-Dev: [email protected]
http://lists.linux-ha.org/mailman/listinfo/linux-ha-dev
Home Page: http://linux-ha.org/

Reply via email to