In this article
Tailscale netmap caching keeps your tailnet map on disk
Tailscale has introduced netmap caching, a client feature that saves a copy of the tailnet's network map (the netmap) to disk. The netmap holds settings for the node itself, such as DNS, routes and capabilities, plus information about other nodes, including their endpoints and home DERP region.
Tailscale says the cache starts the client faster and reaches nodes faster, especially on non-ideal connections. It also says the client connects to some or all nodes at startup even when the coordination server is unavailable.
- Relayed access after boot: a computer can usually reach other tailnet nodes over an encrypted WireGuard connection relayed via DERP, even if the coordination server is unreachable.
- No direct connections without the server: direct connections are not possible while the coordination server is down, because peers cannot decrypt DISCO messages until they learn your node's new public key.
- Same home DERP as before: the cached self-node data lets a node pick its previous DERP as home instead of the nearest one, which helps other nodes find it.
- Fewer disk writes: the cache is split into chunks, such as the device itself and each other device, and only changed chunks are written again to limit storage wear.
- Policy-scoped contents: the cache holds nothing about nodes you cannot connect to, because the coordination server only reports nodes your tailnet policies allow.
- Self-cleaning: when a full netmap arrives, the cache is updated and old items the server did not explicitly delete are expired.
Why a node loses its netmap after a restart
Without the cache, a node keeps netmap data only in memory while Tailscale runs. New connections still work during a temporary coordination server outage, and established connections are unaffected. A restart or profile switch wipes that memory, and that is the gap the cache fills.
Each node's netmap is unique, since self-node data differs and policy-file rules on who may talk to whom are enforced through it. After sending a full netmap, the coordination server sends only changes. Endpoints change often, while DNS and other settings stay fairly static.
Netmap caching is opt-in and absent on mobile
Netmap caching needs an optional flag to activate. Tailscale says it could become a default feature after successful testing. You can enable, disable and manage it through the tailnet policy file.
It is not available on Android and iOS yet. Tailscale says it will come later and gives no date.
Try Tailscale netmap caching on flaky links, no rush
This is a routine, opt-in feature, so nothing here is urgent. If you run nodes on unreliable links or nodes that reboot while the coordination server may be unreachable, it is worth testing. To report problems, contact Tailscale support or open a GitHub issue with the label client-side-caching. The announcement is the Tailscale blog post listed under Sources.
What Tailscale has not published about netmap caching
The announcement does not say which client release includes the feature or when it ships. It does not name the flag or policy setting that turns the cache on, and it offers no startup-time measurements.
Sources
Image: Image by VISHNU_KV from Pixabay
This story was produced with AI assistance from the sources listed above and fact-checked against them before publication. Read our editorial methodology and the corrections log. Spotted an error? Email [email protected].


Discussion