| 1 | package host |
| 2 | |
| 3 | import ( |
| 4 | "errors" |
| 5 | "fmt" |
| 6 | "github.com/carlosframework/platform/internal/registry" |
| 7 | "log" |
| 8 | "path/filepath" |
| 9 | "time" |
| 10 | ) |
| 11 | |
| 12 | // Frozen from origin/main 34a4328f, internal/host/agent.go. Only names, |
| 13 | // clock and readiness injection differ, so the oracle cannot drift with |
| 14 | // the production legacy helpers. |
| 15 | func fix5MainCheckedRestart(sys unitSystem, optDir string, getenv func(string) string, settled func(rt registry.Route), c rootRestartChecks) func(rt registry.Route) error { |
| 16 | return func(rt registry.Route) error { |
| 17 | before, err := sys.Show(rt.Unit) |
| 18 | if err != nil { |
| 19 | log.Printf("adopt: %s: cannot read %s (%v); restart not checked", rt.Host, rt.Unit, err) |
| 20 | return sys.TryRestart(rt.Unit) |
| 21 | } |
| 22 | link := filepath.Join(optDir, "live", rt.Host) |
| 23 | if before.ExecPath != link { |
| 24 | log.Printf("adopt: binding: %s runs %s, but the unit stamp is on %s; restart not checked", rt.Unit, before.ExecPath, rt.Host) |
| 25 | return sys.TryRestart(rt.Unit) |
| 26 | } |
| 27 | if before.ActiveState != "active" { |
| 28 | return sys.TryRestart(rt.Unit) // stopped by an operator: stays a no-op |
| 29 | } |
| 30 | want, err := filepath.EvalSymlinks(link) |
| 31 | if err != nil { |
| 32 | return sys.TryRestart(rt.Unit) |
| 33 | } |
| 34 | if err := sys.TryRestart(rt.Unit); err != nil { |
| 35 | return err |
| 36 | } |
| 37 | timeout := unitSettleTimeout(getenv) |
| 38 | err = settleUnit(sys, rt.Unit, before.InvocationID, settleOpts{ |
| 39 | Timeout: timeout, Poll: 500 * time.Millisecond, Hold: 5 * time.Second, |
| 40 | WantDir: filepath.Dir(want), Ready: c.legacyReady(rt), |
| 41 | Now: c.now, Sleep: c.sleep}) |
| 42 | if errors.Is(err, errUnitStopped) { |
| 43 | // An operator stopped it between the read above and now. Not a |
| 44 | // verdict on the build, and a rollback restart would undo their |
| 45 | // stop: same as a unit that was stopped before the restart. |
| 46 | log.Printf("adopt: %s: %s was stopped during settle; not judged, not restarted", rt.Host, rt.Unit) |
| 47 | return nil |
| 48 | } |
| 49 | if err == nil { |
| 50 | log.Printf("adopt: %s: %s settled on %s", rt.Host, rt.Unit, filepath.Dir(want)) |
| 51 | // D3 (proofobserve.go): a build that passed the settle check is a |
| 52 | // build this box has run well. |
| 53 | if settled != nil { |
| 54 | settled(rt) |
| 55 | } |
| 56 | } |
| 57 | return err |
| 58 | } |
| 59 | } |
| 60 | |
| 61 | // rollbackUnitRestart is D1's rollback restart (release.WithRollbackRestart): |
| 62 | // the live link already names the restored build; replace whatever is |
| 63 | // running (a deadline failure can leave the rejected build up) and check |
| 64 | // the restored build comes up too. |
| 65 | func fix5MainRollbackRestart(reg *registry.Registry, sys unitSystem, optDir string, getenv func(string) string, c rootRestartChecks) func(host string) error { |
| 66 | return func(host string) error { |
| 67 | rt, err := reg.Get(host) |
| 68 | if err != nil || rt == nil || rt.Unit == "" { |
| 69 | return nil |
| 70 | } |
| 71 | before, _ := sys.Show(rt.Unit) |
| 72 | if err := sys.ResetFailed(rt.Unit); err != nil { |
| 73 | log.Printf("adopt: %s: reset-failed %s: %v (continuing)", host, rt.Unit, err) |
| 74 | } |
| 75 | if err := sys.Restart(rt.Unit); err != nil { |
| 76 | return err |
| 77 | } |
| 78 | want, err := filepath.EvalSymlinks(filepath.Join(optDir, "live", host)) |
| 79 | if err != nil { |
| 80 | return err |
| 81 | } |
| 82 | err = settleUnit(sys, rt.Unit, before.InvocationID, settleOpts{ |
| 83 | Timeout: unitSettleTimeout(getenv), Poll: 500 * time.Millisecond, Hold: 5 * time.Second, |
| 84 | WantDir: filepath.Dir(want), Ready: c.legacyReady(*rt), |
| 85 | Now: c.now, Sleep: c.sleep}) |
| 86 | if errors.Is(err, errUnitStopped) { |
| 87 | // The restored build is what the link names; a stop an operator |
| 88 | // made while it settled is theirs to keep, not a failed recovery. |
| 89 | log.Printf("adopt: %s: %s was stopped during the rollback's settle; not judged, not restarted", host, rt.Unit) |
| 90 | return nil |
| 91 | } |
| 92 | if err != nil { |
| 93 | return fmt.Errorf("rollback did not come up either: %w", err) |
| 94 | } |
| 95 | return nil |
| 96 | } |
| 97 | } |
| 98 | |
amadan