fix: grain up handles existing/stale daemon correctly by cxdy · Pull Request #12 · cxdy/grain

cxdy · 2026-07-27T06:58:18Z

Summary

grain up could print success while an old daemon still owned the socket/API, or after a new child failed to bind. Users saw daemon not up right after a successful-looking grain up.

Changes

Detect already-healthy daemon → print already up
Refuse second start if pid is live but unhealthy
Clean stale pid/socket before start
Wait for healthz or child exit (no false success)
Setsid on background daemon
Harder grain down cleanup

Test plan

go test ./internal/cli/
On Linux: leave a stray grain up --fg, run grain up → should say already up or tell you to grain down
Kill daemon uncleanly, leave socket; grain up should clean and start

Previously grain up always spawned a new process and treated any existing socket as success, so a leftover grain up --fg could hold the API port while a second up reported a dead pid. Now: - healthy daemon → "already up" (no second process) - live pid but unhealthy → error pointing at grain down - dead pid / orphan socket → cleaned before start - background start waits for healthz or child exit (no false success) - Setsid so shell Ctrl+C does not signal the daemon - grain down cleans stale files and SIGKILLs if needed

github-actions · 2026-07-27T06:59:40Z

Coverage (lines): 74% ❌ — minimum 75%

_{commit 4971a9a · 9383/12627 lines · cmd/* and tray excluded}

Coverage report by file (74% overall — click to expand)

File	Coverage		Missing
All files	`74%`	❌
`internal/cli/uninstall.go`	`22%`	❌	31-37 45-49 51-57 59-67 69-78 80-82 85-92 95-114 116 118-122 133-135 137-139 144-153 15…
`internal/netutil/port.go`	`45%`	❌	11-13 16-18
`internal/agent/dial.go`	`46%`	❌	44-46 61-63 77-78 83 85-102 106-113
`internal/sshkey/sshkey.go`	`53%`	❌	18-20 33-35 38-44 46-48 50-52
`internal/hypervisor/disk_local.go`	`56%`	❌	38-40 45-47 52-54 56-60 62-69 73-76 78-80 96-98 112-114 123
`internal/cloudinit/seed.go`	`57%`	❌	46-48 51-53 57-65 77-89 101-103
`internal/cli/fs.go`	`63%`	❌	52-54 56-58 60-62 85-101 123-125 127-129 131-133 144-153 175-177 179-181 183-185 191-19…
`internal/cli/stats.go`	`64%`	❌	19-21 23-25 27-29 40-42
`internal/cli/cp.go`	`64%`	❌	58-60 62-64 72-74 80-82 88 136-139 141-143 145 155-157 169-171 179-181 188-190 200-202 …
`internal/cli/tray.go`	`65%`	❌	26-43 58-62
`internal/cli/root.go`	`65%`	❌	135-137 144-147 157-159 167-169 182 185-190 194-204 206 208-210 223-225 241-247 249-252…
`internal/hypervisor/firecracker.go`	`66%`	❌	139-141 153-163 165-170 173 177-202 204-210 213-236 239-243 246-247 250-251 253-257 262…
`internal/api/client.go`	`67%`	❌	46-48 66-68 70-72 74-76 79-81 87-89 97-99 101-103 105-107 111-113 154-156 169-171 173-1…
`internal/cli/secret.go`	`69%`	❌	43-45 47-49 52-54 56-58 66-68 103-105 107-109 112-114 120-122 140-142 144-146 149-154 1…
`internal/agent/client.go`	`69%`	❌	55-57 59-61 77-79 102-104 123-125 127-129 132-134 155-157 164-172 176-178 180-182 221-2…
`client/client.go`	`70%`	❌	96-98 105-107 122-124 126-128 143-145 147-149 164-166 186-188 190-192 194-196 199-201 2…
`internal/cli/proxy.go`	`70%`	❌	43-45 79-86 88-99 101-105 107 109-114 116 129-131 133-135 141-143 155-159 164 174-176 1…
`internal/cli/image_doctor.go`	`71%`	❌	34-38 53-58 62-71 74-76 102-104 139-141 150-152 157 161-163 174-176 181-187 190-192 203…
`internal/proxy/store.go`	`73%`	❌	34-36 61 64-66 73-75 77-79 88-90 110-112 121-123 165 168-171 177-179 181-183 192-194 19…
`internal/store/store.go`	`73%`	❌	21-23 39-41 43-45 47-49 61 74-78 107-109
`internal/image/catalog.go`	`74%`	❌	69-79 119-132
`internal/agent/server.go`	`74%`	❌	91-93 95-97 103 204-206 293-299 332-334 338-341 343-346 388-397 409-421 471-474 476-479…
`internal/cli/fwd.go`	`76%`	✅	46-48 127-129 131-133 136-138 143-145 149-151 189-192 207-209 211-213 221-226 233-235 2…
`internal/presets/presets.go`	`76%`	✅	31-33 40-42 45-46 49-50 62-64
`internal/secrets/store.go`	`76%`	✅	58-60 84-88 101 119-121 148-150 153-159 171-173 197-203 217-219 235 247-249 259 273-275…
`internal/cli/logs.go`	`77%`	✅	37-39 44-46 52 55-59 84-86 104-109 113 123 127-128 141-143 158-160 182 209-211 219-221
`internal/api/api.go`	`77%`	✅	112-115 215-219 228-251 273-275 398-401 412-415 427-429 437-439 500-503 512-514 519-522…
`internal/manager/manager.go`	`78%`	✅	86-88 183-185 201-205 207-209 215-217 238-240 252-258 263-268 380-382 430-432 440-442 4…
`internal/hypervisor/qemu.go`	`78%`	✅	33-35 50-52 66-68 72-74 79-81 88-90 95-97 155-162 168-170 174-184 186-189 196-198 202-2…
`internal/cli/progress.go`	`78%`	✅	43-49 51-54 63-65 103-111 113-116 131-133 195-197
`internal/image/pull.go`	`78%`	✅	106-108 136 157-159 170-172 175-177 188-190 201-205 214-218 220-223 237-239 262-264 267…
`internal/proxy/server.go`	`79%`	✅	65-67 114-118 124-127 168-170 173-175 178-181 191-194 198-202 205-208 222-224 227-233 2…
`internal/hypervisor/virtiofs.go`	`80%`	✅	28-30 32-34 64-66 104-106 111-114 119-121 126-129 170-172 190-192 199-201
`internal/cli/volume.go`	`82%`	✅	24-29
`internal/daemon/daemon.go`	`82%`	✅	27-29 57-60 88-90 98-100 105-107
`internal/guest/sshwait.go`	`83%`	✅	32-34 39-41 58
`internal/hypervisor/fsdev.go`	`83%`	✅	34-40
`internal/agent/stats.go`	`83%`	✅	62-64 71-73 80-82 116-118 125-127
`internal/hypervisor/qmp.go`	`84%`	✅	50-52 62-69 93-95 97-99
`internal/cli/act.go`	`87%`	✅	97-99 101-103 112-116 119-121 143-145 150-152 166-168 209-211 218-220 254-256 337-339
`internal/guest/agentdeploy.go`	`87%`	✅	59-61 65-67 75-77
`internal/hypervisor/transport.go`	`91%`	✅	69-71
`internal/agent/binary.go`	`91%`	✅	58-59 65-66
`internal/hostbin/hostbin.go`	`91%`	✅	28-30
`internal/config/config.go`	`92%`	✅	306-308 326-328 344-346 385 396-398 426-428 435-437
`internal/cli/preset.go`	`92%`	✅	36-38
`internal/hypervisor/netdev.go`	`93%`	✅	43-45
`internal/cli/agent_dial.go`	`95%`	✅	44-48 150 159
`internal/cli/client.go`	`95%`	✅	39-41
`internal/cloudinit/merge.go`	`96%`	✅	124-126 180-182 238-240
`internal/cli/profile.go`	`97%`	✅	28-30
`internal/hypervisor/mock.go`	`97%`	✅	161-163

Stop the local daemon, remove the CLI binary, and optionally purge the data directory with --purge (-y skips the prompt). Default keeps VMs and images; always drops the installer agent cache and runtime files.

Same-repo PRs already post coverage from ci.yml and never upload the coverage-report artifact. The workflow_run job still ran for every PR and failed with "Artifact not found". Gate on head_repository != base and use the compact coverage-pr-comment.sh script for fork posts too.

cxdy added 2 commits July 27, 2026 03:01

feat: add grain uninstall (and document it)

c2cd994

Stop the local daemon, remove the CLI binary, and optionally purge the data directory with --purge (-y skips the prompt). Default keeps VMs and images; always drops the installer agent cache and runtime files.

cxdy merged commit 0e5a52d into main Jul 27, 2026
1 check passed

cxdy deleted the fix/grain-up-stale-daemon branch July 27, 2026 07:30

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

fix: grain up handles existing/stale daemon correctly - #12

fix: grain up handles existing/stale daemon correctly#12
cxdy merged 3 commits into
mainfrom
fix/grain-up-stale-daemon

cxdy commented Jul 27, 2026

Uh oh!

github-actions Bot commented Jul 27, 2026 •

edited

Loading

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

1 participant

Conversation

cxdy commented Jul 27, 2026

Summary

Changes

Test plan

Uh oh!

github-actions Bot commented Jul 27, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

1 participant

github-actions Bot commented Jul 27, 2026 •

edited

Loading