Files
mesh-controller/internal/inventory/plans.go
T
jochen 4ac5cfe3a7 Roll a plan's module out to one machine first (hq issue 249, ADR 0218)
A plan sent every machine running a module at once, ignoring the module's
upgrade policy. Unless the policy says together, the first machine by name
is sent, recorded in the plan, and the rest follow only once its report
after the send says it applied; a failed first machine stops the plan.
2026-10-05 18:17:52 +02:00

135 lines
4.8 KiB
Go

package inventory
import (
"context"
"encoding/json"
"errors"
"fmt"
"time"
"github.com/jackc/pgx/v5"
)
// A Plan is what a merge produces (novox/hq ADR 0162): the modules it changed and everything
// standing on them, sorted into tiers, each module's state, and the tier the plan is at. Kept in
// the store so a controller replaced mid-plan resumes it, and so `status` can say what a merge
// still waits for.
type Plan struct {
ID string `json:"id"`
Repository string `json:"repository"`
Commit string `json:"commit"`
Created time.Time `json:"created"`
Updated time.Time `json:"updated"`
State string `json:"state"`
Tier int `json:"tier"`
Tiers [][]string `json:"tiers"`
Modules map[string]*PlanModule `json:"modules"`
Note string `json:"note,omitempty"`
}
// PlanModule is one module's state within a plan.
type PlanModule struct {
// State: asked, built, failed; empty for a module whose tier has not been asked yet.
State string `json:"state,omitempty"`
AskedAt *time.Time `json:"asked_at,omitempty"`
BuiltAt *time.Time `json:"built_at,omitempty"`
// SentAt is when the plan sent the machines running this module its new build, because a
// later tier is built by it (ADR 0163's gate): the reports that open the gate are the ones
// after this.
SentAt *time.Time `json:"sent_at,omitempty"`
// First is the machines the plan sent the new build to first, and FirstAt when (novox/hq issue
// 249, ADR 0218): unless the module's policy rolls it out together, one machine takes it before
// the rest, and the rest are sent once that one reports it applied. Kept so a controller
// replaced while the plan waits on that report resumes the wait rather than sending again. The
// machine holding the bus is among them when its user list had to go first.
First []string `json:"first,omitempty"`
FirstAt *time.Time `json:"first_at,omitempty"`
Commit string `json:"commit,omitempty"`
Why string `json:"why,omitempty"`
}
// The states a plan passes through.
const (
PlanBuilding = "building"
PlanRolling = "rolling"
PlanDone = "done"
PlanFailed = "failed"
)
// Open says whether the plan is still being worked.
func (p Plan) Open() bool { return p.State == PlanBuilding || p.State == PlanRolling }
// SavePlan writes a plan, new or changed, whole: the plan is small and read as one thing.
func (i *Inventory) SavePlan(ctx context.Context, p Plan) error {
tiers, err := json.Marshal(p.Tiers)
if err != nil {
return err
}
modules, err := json.Marshal(p.Modules)
if err != nil {
return err
}
_, err = i.store.Pool().Exec(ctx,
`insert into release_plan (id, repository, commit_hash, created, updated, state, tier, tiers, modules, note)
values ($1, $2, $3, $4, now(), $5, $6, $7, $8, $9)
on conflict (id) do update set updated = now(), state = excluded.state, tier = excluded.tier,
tiers = excluded.tiers, modules = excluded.modules, note = excluded.note`,
p.ID, p.Repository, p.Commit, p.Created, p.State, p.Tier, tiers, modules, p.Note)
return err
}
// OpenPlans is every plan still being worked, oldest first.
func (i *Inventory) OpenPlans(ctx context.Context) ([]Plan, error) {
return i.plans(ctx, `where state in ('building', 'rolling') order by created`)
}
// RecentPlans is the last few plans, newest first, open or not — what the overview shows.
func (i *Inventory) RecentPlans(ctx context.Context, limit int) ([]Plan, error) {
return i.plans(ctx, fmt.Sprintf(`order by created desc limit %d`, limit))
}
// PlanByID is one plan.
func (i *Inventory) PlanByID(ctx context.Context, id string) (Plan, error) {
plans, err := i.plans(ctx, `where id = '`+id+`'`)
if err != nil {
return Plan{}, err
}
if len(plans) == 0 {
return Plan{}, fmt.Errorf("no plan %s", id)
}
return plans[0], nil
}
func (i *Inventory) plans(ctx context.Context, tail string) ([]Plan, error) {
rows, err := i.store.Pool().Query(ctx,
`select id, repository, commit_hash, created, updated, state, tier, tiers, modules, note
from release_plan `+tail)
if err != nil {
return nil, err
}
defer rows.Close()
var out []Plan
for rows.Next() {
var p Plan
var tiers, modules []byte
if err := rows.Scan(&p.ID, &p.Repository, &p.Commit, &p.Created, &p.Updated, &p.State,
&p.Tier, &tiers, &modules, &p.Note); err != nil {
return nil, err
}
if err := json.Unmarshal(tiers, &p.Tiers); err != nil {
return nil, err
}
if err := json.Unmarshal(modules, &p.Modules); err != nil {
return nil, err
}
if p.Modules == nil {
p.Modules = map[string]*PlanModule{}
}
out = append(out, p)
}
if errors.Is(rows.Err(), pgx.ErrNoRows) {
return nil, nil
}
return out, rows.Err()
}