Docker Daemon Address Pool Planner

Features

• Generates a snippet ready to paste into /etc/docker/daemon.json
• Subnet size can be Auto: pick the smallest prefix that fits the containers per network (network, broadcast and bridge gateway already deducted)
• Lists which subnet each network gets, in order from the pool, with usable addresses, container capacity and first/last usable address
• Checks overlap with docker0, the default pool and the bridge default, and the subnets against each other
• Optionally includes bip (the docker0 bridge address)
• Points out the canonical network address when base carries host bits
• ⚠️ Clear errors when it does not fit: a subnet larger than the pool reports the prefix you need, and too many networks reports which one fails and how wide base must be
• All local, so the config never leaves your browser

How to use

1. Pick the subnet size per network (Auto, or /16 through /28)
2. With Auto, enter the containers per network; otherwise skip it
3. Set how many networks to create and the pool base
4. Tick bip if you also want the docker0 bridge setting
5. Copy the daemon.json snippet or check the subnet list (use the other tool for compose ↔ docker run conversion)

FAQ

Not sure how large the subnet should be? Use Auto

Set Subnet size to Auto and give the containers per network; the tool picks the smallest prefix that fits. ⚠️ The arithmetic is not simply rounding the container count up to a power of two: every subnet gives up its network and broadcast addresses, and a Docker bridge takes one more as its gateway, so the capacity is 2^(32 − prefix) − 3. A /24 holds 253 containers, not 256, so 254 containers needs a /23. ⚠️ The auto prefix never goes past /30 — Docker rejects /31 and /32 as a pool size.

What is the difference between default-address-pools and bip?

bip sets the address of the default bridge docker0 itself (172.17.0.1/16 out of the box), while default-address-pools controls where user-defined bridge networks created by docker network create are cut from and how large they are. The two must not overlap, and Docker checks this at startup.

Why is base /16 with size /28 not an error?

Because cutting /28 subnets from a /16 pool is perfectly valid (up to 4096 of them); the official example uses base /16 with size /24. The genuinely wrong case is the reverse: a size prefix smaller than base (a subnet larger than the pool), or a network count that exceeds what the pool can yield. This tool errors on those and merely warns about thousands of subnets.

Why does the default 172.17.0.0/16 report a conflict?

Because that is exactly the range docker0 occupies out of the box, and it also sits inside Docker default pool 172.16.0.0/12. To avoid it, choose a base inside 10.0.0.0/8, for example 10.201.0.0/16.