Repository navigation
vm: kopia metadanych obrazu na Drive, alarm limitu Google, pomiar zajętości, docs tapu - #34
Merged
Merged
Conversation
…tale The bands on Drive cannot be opened without the image's Info.plist, token and mapped/, which lived only on the machine's disk - kept out of Time Machine. tm-meta-sync (service + 15 min timer) copies /srv/tm-local minus bands/ and lock to <folder>-tm-meta, never from an image without its Info.plist (rclone sync would wipe the copy), and stamps each success. vm-status shows the copy's age; it and backup-health report one older than an hour (deferred after startup). vm/README gains the restore procedure. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01AU1MhxJW1e9oTB2XvjX8WT
…sured The 14:31 reading (101 MB, before the move to <folder>-tm) was still shown at 18:00 while the folder held 215 GiB: younger than six hours, it was never measured again. Readings now carry their folder; another folder's reading, or one stored before this change, is measured again instead of shown. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01AU1MhxJW1e9oTB2XvjX8WT
renacode-tap is on the bypass list of homebrew-tap's main ruleset since 2026-10-27 and 1.3.15 reached the tap on its own. The hand-made pull request is now only the workaround for a failed push. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01AU1MhxJW1e9oTB2XvjX8WT
At the 750 GB limit rclone's write-back retries every upload, at most 5 min apart, and resumes by itself (rclone 1.75.1 vfs/vfscache/writeback) - but nothing reported it: erroredFiles counts cache cleaning, not uploads, and Time Machine kept finishing into the machine's buffer, so every light stayed green while nothing reached Drive. The status reads the end of rclone's log: a refusal newer than the last upload means the limit holds. vm-status and backup-health report it. vm-guard --dry-run prints the guard's decision for every Time Machine state instead of acting, so thresholds can be tried on a throwaway machine without pausing the live backup (Time Machine is global). Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01AU1MhxJW1e9oTB2XvjX8WT
On 27.10.2026 a new machine's first boot took 30 min: Lima copies every `mode: data` file with `sudo -iu root`, and while cloud-final runs logind cannot open the session, so pam_systemd times out after 2 min per file. `limactl start` gave up after its default 10 min and vm-setup failed on a new Mac or a restore. Later boots take ~12 s. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01AU1MhxJW1e9oTB2XvjX8WT
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Kontynuacja #32/#33 (sesja 27.10.2026, wieczór).
Co
vm/systemd/tm-meta-sync.service+.timer, co 15 min, w szablonie):rclone sync /srv/tm-local/ ${CM_REMOTE}-meta/bez*/bands/**i*/lock. Bez tego Info.plist/token/mapped/obrazu były tylko na dysku maszyny (wykluczonym z TM) i pasm na Drive nie dało się otworzyć.ExecConditionnie puszcza synchronizacji z obrazu bez Info.plist (nowa maszyna przed odtworzeniem skasowałaby kopię). Znacznik/var/lib/cloudmachine/tm-meta-sync.okpo każdym sukcesie.VMLayout.services+tm-meta-sync.timer(unitName()- vm-setup startował<nazwa>.service),vm-statuspokazuje wiek kopii metadanych,vm-statusibackup-healthalarmują po >1 h (odroczone w oknie startu).erroredFilesdotyczy czyszczenia cache), a TM kończył kopie „z sukcesem” do bufora. Status czyta koniec logu rclone: odmowa nowsza niż ostatnia udana wysyłka = limit trwa.vm-status+backup-health(drive-upload-limit).vm-guard --dry-run: decyzje strażnika dla każdego stanu TM bez działania (TM jest globalny - strażnik nacmtestzatrzymałby żywą kopię).-tm) jest mierzony od nowa zamiast pokazywany obok 215 GiB.vm-setup:limactl start --timeout 50m- pierwszy rozruch nowej maszyny trwał 30 min (sudo -iu rootLimy przy każdym plikumode: data, pam_systemd czeka 2 min w trakcie cloud-final), domyślne 10 min kończyło vm-setup błędem. Kolejne rozruchy ~12 s.renacode-tapod 27.10, 1.3.15 weszło samo); ręczny PR tylko jako obejście.Testy
swift test: 305 zielonych;swift-format lint --strictczysto.systemd-analyze verifyjednostek w maszynie: OK.cmtest(gdrive:CloudMachine/test-vm-tm) z tej gałęzi:-tm-meta, znacznik;vm-statusstany świeża / ZALEGŁA (rc 1) / NIGDY;backup-health --quietzgłasza i gasi alarm.limactl delete cmtest, nowa maszyna, kroki z README,hdiutil attach -readonly- SHA-256 pliku identyczne.Received upload limit errorw logu →vm-status/backup-healthzgłaszają, po udanej wysyłce gaśnie. Odcięcie 443 w maszynie: kolejka 47 pasm rośnie,--dry-run= pause; po przywróceniu sieci rclone dosłał wszystko sam w ~1 min bez restartu,--dry-run= resume. Uwaga: odcięta sieć to nie 403 Google - 8 wysyłek wisiało 47 min na zerwanych połączeniach (do osobnej analizy).Poza zakresem
cmsmbto prototyp (jednostki na sztywno, bezvm.env/bundle.env, bezsmbd.service.d/cloudmachine.conf) - przepięcie osobno, za zgodą.~/.lima/<maszyna>/lima.yamlprzy tworzeniu; nowy szablon z wydania NIE trafia do istniejącej maszyny (komentarz wcmsmb.yamlmówi inaczej) - do osobnego PR (vm-update).🤖 Generated with Claude Code
https://claude.ai/code/session_01AU1MhxJW1e9oTB2XvjX8WT