--- title: "SOA Settings" description: "The zone-level SOA record: hostmaster contact, refresh, retry and expire for secondary nameservers, and the minimum TTL that governs negative caching." category: "Deployment" language: en source: https://cubepath.com/docs/products/deployment/dns-soa-settings updated: 2026-11-17 --- # SOA Settings Open the zone → **SOA Settings** tab in [my.cubepath.com](https://my.cubepath.com). The SOA (Start of Authority) is the record that describes the zone itself rather than any name in it: who's authoritative for it, who to contact about it, and the timers that secondary nameservers (servers that keep a copy of your zone) follow. Every zone has exactly one, created with the zone. Most zones never need this tab. It matters when you're moving a domain in from another provider and want the timers to match, when you run your own secondary nameservers, or when your compliance paperwork asks who the hostmaster is. ## Read-only fields | Field | What it is | |-------|------------| | **Primary nameserver** | The nameserver named as authoritative in the SOA. We set and maintain it | | **Serial** | The zone's version number. It goes up on its own every time the zone changes; secondaries use it to notice they're out of date | ## Editable fields | Field | Default | Range | What it does | |-------|---------|-------|--------------| | **Hostmaster email** | `hostmaster@` your domain | n/a | The contact address for zone problems, e.g. `hostmaster@example.com`. It's published in the SOA, so use a role address, not a personal one | | **Refresh** | 3600 s | 300–86400 | How often a secondary checks the serial for changes | | **Retry** | 900 s | 60–86400 | How long a secondary waits before retrying a refresh that failed | | **Expire** | 604800 s | 86400–2419200 | How long a secondary keeps serving the zone when it can't reach the primary. After this it stops answering | | **Minimum TTL** | 300 s | 60–86400 | The **negative caching** TTL: how long resolvers cache the fact that a name *doesn't exist* | Change what you need and click save; the serial advances and secondaries pick the change up on their next refresh. ## Minimum TTL is the one that bites The name is misleading: it isn't a floor on your records' TTLs, and it doesn't affect answers that exist. It controls how long a resolver remembers an **NXDOMAIN**: a query for a name that isn't in your zone. That matters when you add a record people have already tried to reach. If someone hits `api.example.com` before it exists, their resolver caches the "no such name" for the minimum TTL, and adding the record doesn't help them until it expires. If you're about to publish new names and expect early traffic, lower this first. > Note: 300 seconds is a good default and a deliberately modest one. Very high negative caching turns a typo you fixed in seconds into an outage that lasts hours for the people who hit it first. ## Refresh, retry and expire These three only mean anything to a **secondary** nameserver, one that copies the zone from the primary. If CubePath's nameservers are the only ones serving your domain, they change nothing about how visitors resolve it, and the defaults are fine forever. If you do run secondaries, the shape to aim for is: refresh often enough that changes propagate on a timescale you're happy with, retry much faster than that so a brief failure doesn't stall you, and expire long enough that a multi-day outage of the primary doesn't take the zone off the internet. The defaults (one hour, fifteen minutes, one week) are that shape already. > Warning: **Expire** is the dangerous one. Setting it low means a secondary stops answering for your domain after a comparatively short loss of contact with the primary. Leave it in days, not hours. ## Related - The records themselves, and per-record TTLs: [Records](/docs/products/deployment/dns-records). - Zone creation, nameserver delegation and verification: [DNS Zones](/docs/products/deployment/dns-zones).