MetronIPTV
Database · primary
High availability
Two servers. One address. No downtime during maintenance.
The Emiton HA cluster keeps Metron and IPTV on two nodes. The database replicates continuously, the shared viewer address is taken over by the second node on failure, and updates roll out one server at a time.
Your IPTV and Metron keep running even when one server goes down — no technician on the line, no reconfiguration for your viewers.
2 nodes · active + standby ~20–25 s database failover synchronous replication no lost committed writes no extra hardware for arbiters
How it works
One cluster, two equal nodes
Viewers · TV · apps
Shared address (VIP) iptv.your-domain.tld
MetronIPTV
Database · replica
Arbiter Arbiter Arbiter an odd number of members votes on who leads — reusing existing streamers or the licence server
Viewers always reach a single shared address, never a server directly. On failure it moves to the healthy node; the database keeps a synchronous copy on both sides.
- 01 Two nodes run at once
The leader serves your viewers; the second keeps a continuous copy of the database.
- 02 Arbiters watch the health
When the leader stops answering, a majority vote elects a new leader.
- 03 The address moves
Viewers stay on the very same address — it is simply served by the other server now.
In detail
What the cluster gives you
Protected
What keeps running
Metron (CRM, billing, customers) and IPTV (logins, playlists, EPG). Stream transcoding stays on the streamers — the cluster does not carry that load.
Failover
When a node goes down
The shared address is taken over by the second node within a few seconds. The database promotes a new primary in about 20–25 seconds, with no loss of committed writes.
Updates
Upgrades without downtime
A controlled update rolls the nodes one at a time. The new version is verified against a database copy first, and only then takes over live traffic.
Quorum
Arbiters decide the leader
An odd number of members votes on which node leads (up to five). An arbiter can be an existing streamer or the licence server — no extra hardware.
Security
Closed between the nodes
The firewall between nodes configures itself; cluster ports are open only to its members. Certificates are shared encrypted and reconciled every 10 minutes.
Requirements
What you need
Two servers on the same L2 network, one shared IP address, synchronised time and an HA-pair licence. Moving an existing install over takes a single restart.
Verified parameters
The numbers, as they really are
We don't promise “zero downtime” or “instant database failover”. These are the values the cluster actually holds.
| Parameter | Value | Note |
|---|---|---|
| Main nodes | 2 | HA pair — licence “Nodes (HA) = HA pair” |
| Signal between nodes (VRRP) | 1 s | application health check every 2 s |
| New database primary | ~20–25 s | Patroni TTL 20 s, 26 s in testing |
| Database replication | synchronous | committed writes are never lost |
| Voting members | 3 or 5 | always an odd number, at most 5 |
| Certificate reconcile | 10 min | and immediately after a new one is issued |
| Protected modules | Metron · IPTV | transcoding stays on the streamers |
Questions
Frequently asked
Do the channels I'm watching get interrupted?
On failover the player reconnects to the same address within a few seconds.
Do I need new hardware?
A second server, yes — it can be a VM. Arbiters need nothing extra; existing streamers or the licence server are used.
Can I switch from a single server without reinstalling?
Yes. The second server can be a clone of the existing one, and the switch takes a single restart.
What if an arbiter fails?
With five votes the cluster keeps working even without two of them.
Ready to run on two servers?
Keep IPTV and Metron up — through outages and through updates.
The HA cluster pairs your existing install with a second node and a shared address, so a failure or a maintenance window stops being an outage for your viewers.