mix/selfpost
1
0
Fork 0
Code Issues Pull Requests Actions 21 Packages Projects Releases Wiki Activity
Files
60c7b85f84982d119561a21b384ee9d71a1de71e
selfpost/docs/schema-migrations.md
T
mixandClaude Opus 5.5 60c7b85f84
test / test (push) Canceled after 0s
Details
findings of the reviews of audit stage 1 
Technical review (Opus) and security review (Fable) of 9efc9c7..a59e3d4;
reports in the session's scratchpad. Fixed:

- B1 (verified bug): store.UpdateStatus read before writing in its
  transaction; a commit by another connection in between (the dmarc-ingest
  process opens the same file) failed it at once with "database is locked"
  and the delivery line was lost. The first statement is now a write.
- R3: a delivery line updated every row of its queue id and recipient; it
  now updates only the newest (Postfix reuses short ids). The sweep closes
  by row id (QueuedDelivery.ID, CloseUnreported(id)).
- R4: a contradicted sweep keeps a mark, reason "reconcile-contradicted",
  instead of losing which row it was.
- R1: contradicted cannot see a lost line (never read late). New counter
  reconcile_missed: at close, the sweep looks for the row's delivery line
  in mail.log and mail.log.1 (QueueLines). 0007 edited - it has run on no
  database but throwaway stands.
- R2: queued_over_lifetime cannot measure "what removing the sweep would
  leave" while the sweep runs; its text now says rows stuck despite the
  sweep, and the rule uses closed - missed - contradicted.
- cert-reload (R5, R6, R7, security medium/low): both files copied to .new
  and the copies checked as a pair before both are renamed; acts on the copy
  only when postconf says Postfix reads it; a failed reload is retried next
  pass; umask 077; says so when the mounted files cannot be read. The e2e
  case asserts Postfix reads the copy, the copy equals the mount, the key is
  600 root.
- Security low: a recipient with a quoted local part - the journal cut
  "a b"@x at the space, deliveryRe cut "a>b"@x at the ">" - matched no
  delivery line and was closed by the sweep. cleanAddress ends a bracketed
  address at the first ">" outside quotes; deliveryRe takes the recipient
  up to the ">, " before relay= (test fixtures given the relay= every real
  delivery line has).
- Security low: postqueue -p's error carries at most 512 bytes of output.
- Nits: addDeliveryLogDay refuses a column not on its list; the truncation
  test checks its message.
- To stage 2b (page wording, mockup-first): the history step of a bounced
  row says the receiving server refused it, false for a row the sweep
  closed. To the owner when stage 2b is reached: since M13 the "deferred"
  figure counts whole messages too, which "Connections turned away" would
  not name.

Found while verifying: the e2e's main positive path failed once with "553
not owned by user app1" - an smtpd idle from before the panel's reload took
the connection in the reload's second. Measured on the stand (one-off
probe, not committed): sending 0, 10, 40, 110 s after the save (two seconds
at the soonest) always met an smtpd started after the reload. The e2e now
waits for the sender map after each application it creates
(waitForSenderMap: AUTH, MAIL, RCPT, RSET - no message, no row), and the
level-1 case uses it instead of its own loop.

Docs: architecture § Queue reconcile, schema-migrations (rows 7 and 8,
chain head 8), CHANGELOG, audit plan (stage 1 note, the measured window,
a new stage 2b line: 62 lines), roadmap 15/62.

Evidence on the test server for this tree: go vet clean, go test ./... 23
packages ok, make e2e exit 0 (renewed certificate, hung panel, level 1 at
49/60).

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-14 15:21:52 +03:00

152 lines
9.5 KiB
Markdown
Raw Blame History

# SQLite schema and migrations
**What this file is.** A living reference for the embedded SQLite schema: the
2.0 baseline, the migrations added after it, and how a database is opened.
When you add or review a migration, update this file in the same change.
**Source of truth in code:** `internal/store/migrations/*.sql`, applied by
`internal/store/store.go` (`migrate()`).
**Related:** [architecture.md](architecture.md) § Persistence;
[plans/panel-redesign.md](plans/panel-redesign.md) § No compatibility;
[development.md](development.md) § Documentation.
---
## Current snapshot
| Field | Value |
|---|---|
| Chain head | `user_version = 8` |
| Files | `0001_init.sql` (the 2.0 baseline), `0002_inbound_spam_log.sql`, `0003_send_log_lines.sql`, `0004_notifications.sql`, `0005_send_log_rejected_index.sql`, `0006_sender_check.sql`, `0007_delivery_log_days.sql`, `0008_tenant_boundary.sql` |
| File mark | `PRAGMA application_id = 0x53503230` ("SP20") |
| Database file | `/data/selfpost.db` (bind mount) |
| Compatibility | 2.x MINOR releases must boot a `2.0.0` data directory. **Nothing is carried over from 1.x.** |
---
## 2.0 starts from an empty data directory
Decided by the owner on 2026-09-21: the redesign is a breaking release, the
1.x migration chain (`0001` … `0010`, with its `admin` table created in the
first file and dropped in the fifth) is deleted from the binary, and there is
no upgrade path. The only live 1.x installation is the owner's and is set up
again on 2.0. The 1.x chain stays in git history (last present at `1.9.5`).
What that means in practice:
- A fresh `/data` gets the baseline and nothing else.
- **A 1.x database is refused, not migrated.** `store.Open` returns
`ErrForeignSchema` for a file that has a schema version but not the 2.x
`application_id`. Without that check a 1.x file at `user_version = 10`
would read as "already migrated" and the panel would start on tables it
does not know. The file is not modified. The same applies to restoring a
1.x full backup into a 2.0 container.
- A database whose version is higher than the build knows is refused too.
- Domains can still be carried by hand: a 1.x **domain export** (JSON, with
the DKIM key and applications) imports into 2.0 — the transfer format is
not the database.
---
## How migrations run
1. Migrations are embedded at build time (`//go:embed migrations/*.sql`).
2. Filenames are sorted lexicographically; **file order = version number**.
3. `PRAGMA user_version` records progress: after file *i* (0-based), version is
*i + 1*.
4. Each pending migration runs in its own transaction, then bumps `user_version`.
5. There is no down-migration; fixes ship as a new `00NN_*.sql` file.
**Implication:** once `2.0.0` is released, `0001_init.sql` is never edited
again and no file is deleted, renamed or reordered — every 2.x database has
already run it. `2.0.0` and `2.0.1` shipped with the baseline alone; the
features after them add files, the first being `0002_inbound_spam_log.sql` of
inbound-antispam-panel (`2.1.0`), then `0003_send_log_lines.sql` of
delivery-log-storage (`2.2.0`), `0004_notifications.sql` and
`0005_send_log_rejected_index.sql` of panel-notifications (`2.3.0`), and
`0006_sender_check.sql` of from-header-binding (`2.4.0`), and
`0007_delivery_log_days.sql` and `0008_tenant_boundary.sql` of the 2026-10-12
audit (`2.5.0`).
---
## Migration chain
| Ver | File | Shipped | Kind | Summary |
|-----|------|---------|------|---------|
| 1 | `0001_init.sql` | 2.0.0 | DDL | The whole 2.0 schema, below |
| 2 | `0002_inbound_spam_log.sql` | 2.1.0 | DDL | `inbound_spam_log`, `inbound_spam_list` and two indexes; nothing existing changes |
| 3 | `0003_send_log_lines.sql` | 2.2.0 | DDL | `send_log_lines` and one index; nothing existing changes |
| 4 | `0004_notifications.sql` | 2.3.0 | DDL | `users.notify` (default 1: every user starts with notifications on) and `notify_state` |
| 5 | `0005_send_log_rejected_index.sql` | 2.3.0 | DDL | `idx_send_log_rejected`, a partial index of the refused rows of `send_log` by time; a file of its own because the test installation had run 0004 before the review asked for it |
| 6 | `0006_sender_check.sql` | 2.4.0 | DDL | `send_log.reason` and `send_log.header_from` (both default `''`: a row refused earlier keeps no reason), and `sender_check_days` |
| 7 | `0007_delivery_log_days.sql` | 2.5.0 | DDL | `delivery_log_days`: what the reconcile sweep and the log tailer did, a row per UTC day, kept a year, shown nowhere - the measurement that decides the sweep's fate (audit 2026-10-12, stage 1). `send_log.reason` gains the values `reconcile` and `reconcile-contradicted` without a schema change |
| 8 | `0008_tenant_boundary.sql` | 2.5.0 | DDL | `idx_applications_login_nocase`: an application login is unique without regard to case (audit 2026-10-12, H4) |
**Kind:** *DDL* — schema only; *data* — `INSERT`/`UPDATE` that must stay correct
for operators upgrading from older 2.x images.
---
## Schema at head (v2)
| Table | Role |
|-------|------|
| `settings` | Key/value settings of the instance — what is true whoever is signed in (send-log retention). Nothing that belongs to a user |
| `users` | Panel logins. `role` is `global` or `domain`; `email` is the user's own address; `notify` is the user's own switch for notification e-mails (v4); `all_domains` / `all_inbound_domains` widen a `domain` user to every domain of that direction |
| `sessions` | Panel login sessions (token hash, owning user by id and deleted with them, sliding idle expiry) |
| `domains` | Sending (outbound) domains. `dmarc_rua` is where aggregate reports go (`''` = no reports) |
| `user_domains` | Outbound domains assigned to a `domain` user |
| `applications` | SASL applications per domain, with the client-IP restriction (`auth_ip_restrict`, `auth_allowed_ips`) |
| `application_addresses` | Explicit From addresses (`list` mode) |
| `rate_limits` | Level-2 limits per domain / application, manual or auto |
| `send_log` | Delivery journal |
| `logtail_state` | Log-tailer read position |
| `inbound_domains` | Inbound relay domains |
| `inbound_transports` | Upstream host / port / TLS per inbound domain |
| `inbound_recipients` | Allow-list when `recipient_mode = list` |
| `user_inbound_domains` | Inbound domains assigned to a `domain` user — granted separately from the outbound ones |
| `dmarc_reports` | Parsed aggregate report summaries |
| `dmarc_report_records` | Per-source rows inside a report |
| `inbound_spam_log` | Inbound antispam journal: one row per recipient of a message offered on port 25, with the decision (`accept`, `reject`, `tempfail`, `quarantine`) and who made it (`rspamd`, `postfix`, `selfpost`). Metadata only (v2) |
| `inbound_spam_list` | Allow / deny entries for the external filter (`address`, `domain`, `ip`), instance-wide; unique per list (v2) |
| `notify_state` | What the panel has already written to its users about, one row per subject (`dns:<domain>`, `cert`, `queue`, `limit:<domain>`), so that each thing is said once and a restart says nothing again (v4) |
| `send_log_lines` | The `mail.log` lines Postfix wrote under a queue id, with the line's own time; what a message's page shows once the log has rotated. No reference to `send_log`: a line belongs to a row by queue id and time, and goes when no row carries its id (v3) |
`TestBaselineSchema` (`internal/store/schema_test.go`) pins every table, column
and index of this list; `TestOpenMigratesA20Database` opens a database that
holds the baseline alone.
### What changed against the last 1.x schema
| 1.x | 2.0 | Why |
|---|---|---|
| `admin` created in 0001, dropped in 0005 | never exists | the reason for the squash |
| `users.role` = `global` / `domain_admin` | `global` / `domain` | the role is a reach, not a rank — everyone in the panel is an administrator |
| `users.dmarc_report_email` | `users.email` | a user's e-mail is theirs; it is what the panel writes to, and a domain's report address can be filled with it |
| instance setting `dmarc_report_email` (mirror of the global user's field); `domains.dmarc_rua` NULL = inherit the instance default | `domains.dmarc_rua` alone: the address, `''` = no reports | a domain's report address is one of three explicit choices — no reports, hosted, a typed address; nothing is inherited and no user's default exists (2026-10-10) |
| — | `user_inbound_domains` | inbound is delegated per user, separately from outbound |
| — | `users.all_domains`, `users.all_inbound_domains` | "All" includes domains added later |
| `rate_limits.allowed_ips` (unused since 1.9.0) | gone | client IPs belong to the application (`auth_allowed_ips`) |
### Where a domain's DMARC reports go
`domains.dmarc_rua` is the address itself and the only thing stored; `''` means
no reports, and it is what a new domain has. The choice shown in Domain
settings is derived from it: `''` is no reports, the address SelfPost derives
for the domain and the server's hostname (`dmarc.HostedReportAddress`, offered
while hosted reports are enabled) is hosted, anything else is an address typed
for the domain. Nothing refers to a user: deleting one changes no domain.
---
## Adding a migration
1. Add `internal/store/migrations/00NN_short_name.sql` (next number).
2. Update the table in [Migration chain](#migration-chain), the snapshot at the
top, and the affected rows of [Schema at head](#schema-at-head-v2).
3. Update `TestBaselineSchema` if tables, columns or indexes change, and add a
test that opens a database at the previous version and migrates it.
4. A *data* migration states in a comment which 2.x versions it is written for.
Reference in New Issue View Git Blame Copy Permalink