Run machines in a worker pool
A worker pool is vCPU, memory and disk set aside once and shared by the machines in it. On Deployport Compute a machine is a worker in a pool by default, so this is the ordinary way a Compute machine is created.
The pool draws on the account’s funds by the minute while it exists, whether its machines run or not. Its machines draw nothing more.
Create one
Section titled “Create one”In Machinery, open Worker Pools and choose Create. Give it a title and pick a tier.
deployport machinery worker-pools tiersdeployport machinery worker-pools create --title "Dev fleet" --tier <tier>The name is derived from the title, and --name supplies your own instead. Leave --hosting-platform out and the pool is set aside on Deployport Compute, and leave --region out and the first region that platform lists is used.
A worker pool never changes platform afterwards, so name the platform when you create it.
Put a machine in it
Section titled “Put a machine in it”New Machine has a Kind row and a Worker pool row. A worker shows no Place, Region, Zone or Size row, because the pool decides all four.
deployport machinery create --worker-pool <worker-pool-id>The machine shares the pool’s vCPU, memory and disk with the pool’s other machines, and the pool pays for it. Leave --instance-type out with --worker-pool, and leave both out to land in the account’s default pool.
A machine that needs vCPU and memory of its own is a server, created with --kind server, and a machine with a screen is a desktop. Neither joins a pool.
Choose the default pool
Section titled “Choose the default pool”One pool is where a create that names none lands:
deployport machinery worker-pools set-is-default <worker-pool-id> --value trueHow many machines a pool contains
Section titled “How many machines a pool contains”A pool shows the most machines it may contain beside how many it contains now. A 2-vcpu-8gb pool allows up to 50 machines. Only an operator changes that count.
Machines past what the pool’s memory allows slow down while their memory is reclaimed and paged to disk, and in the last case one of them is stopped.
Change the tier, or destroy it
Section titled “Change the tier, or destroy it”deployport machinery worker-pools change-tier <worker-pool-id> <tier>deployport machinery worker-pools destroy <worker-pool-id>A tier change leaves its machines running. Machines already in it keep the vCPU, memory and disk they were given, and machines created after the change get the new tier’s. Destroying a pool gives its vCPU, memory and disk back, so it refuses while machines are in it.
Check it
Section titled “Check it”deployport machinery worker-pools listOne row per pool, newest first. MACHINES is how many machines it contains beside the most it may contain, RUNNING is how many of them run now, and DISK is what its machines have written, their snapshots included, beside the tier’s disk. deployport machinery worker-pools get <worker-pool-id> prints one pool with its swap state and its tier.
Next steps
Section titled “Next steps”- Choose where a machine runs covers the platform, region and zone for a machine that is not in a pool.
- Set what every machine inherits decides the settings its machines run under.
- Launch your first machine creates one with every field left to the account’s defaults.