I am currently busy with a perl script to achieve this, but if there is an existing script to do this even better.ĭate Packet-Type PPP Framed-Protocol User-Name CHAP-Password NAS-Port NAS-Port-Type NAS-Port Connect-Info NAS-Port-Id Service-Type NAS-IP-Address Acct-Session-Id CHAP-Challenge Message-Authenticator Proxy-State I need a script that can parse the flat file to something like this. We have months worth of reply-detail logs, auth-detail logs and detail logs from 4 different servers that needs to be imported.Ī typical reply packet looks as follows - I used a dummy ip for demonstration purposes. Would also like to have an instance of Kibana for basic reporting and overview, and logstash + statsd + Graphite for more sophisticated analysis.I have been searching for an answer to import flat files for radius accounting records into a seperate MYSQL server (not hosted on the same radius server). This main infrastructure in this cluster will consist of elasticsearch+logstash+Redis broker setup and with Graphite for the stats/monitoring (deliveries, bounces, connections, and perhaps a few other metrics such as types of errors, softbounces/hardbounces, etc).
#Parse powermta accounting files install#
The solution involved here will be similar to that described here:įor more info on similar projects see this comment as well: except this will need to be both Linux and Windows PMTA boxes syncing back to the central log processing server (shouldnt matter on the log cluster side, will still be a logstash install on each node environment pushing it back to the central store) We would like the solution to be scalable to over 1000+ systems by simply expanding the Elasticsearch cluster as activity grows. To do this, we need a PMTA Accounting Logs parser written in both Clojure and Python (to compare performance)įor example, we may have 100+ PMTA servers that we'd like to sync to a central log processing server and all logged into a central MySQL DB and a basic report panel written that will show graphs based on the DB info.
We want to be able to automatically deploy logstash on each of the nodes, which pushes all of the PMTA accting log files back to an Elasticsearch cluster, which would be accessed by a dedicated "bounce-processing" server constantly parsing the new logs and storing bounce addresses, deliverable addresses, and other metrics into the centralized database. The goal of this project is to centralize all PMTA delivery accounting files for a large-scale fleet of PowerMTA mail server nodes and to build a central database of non-deliverable (bounced) email addresses and known-good "deliverable" email addresses. Mail Delivery and Automated Parsing and Storage of Bounces and Deliverables.
0 kommentar(er)
