The docs tree is laid out like the menus - a menu path is a docs path. Admin Panel / Settings follows the new structure: Visibility.md is Appearance.md, with new Features.md, Feature-Profiles.md, Security.md, Activities-and-Notifications.md and Delete.md; the section indexes, Admin-Panel.md, Admin-Panel-URLs.md (with the moved-pane redirects) and Validation.md follow the live menus, and Who-Sees-What.md explains in one place which layer decides what is shown and what is allowed. Pages are added for the menu entries that had none, each written from the code: every Admin Panel / Problems pane, the member menu, the board header, Board Settings, the All Boards left menu, and the list, swimlane and card menus. Older links to the moved panes are updated across docs/. TODO.md records what documenting the menus found: Edit Profile does not save a username for non-admins, the Support page's signed-out case is unreachable, Create board from this card is hidden from members who can edit the card, All Boards Settings is unreachable, and the Share template targets live in browser storage. Thanks to xet7 !
2.9 KiB
Admin Panel / Problems / Database problems
What the database said, read for the administrator: which database said it,
what it means, and what to do. Address: /admin/problems/database. Site
administrators only.
Where the rows come from
The database stream of the eventlog collection, written by
server/lib/databaseProblems.js:
- What the database said. Every uncaught error of a method or publication
passes
Meteor._debug; one whose text looks like a database or process resource failure (MongoDB, SQLite, PostgreSQL, MySQL, MariaDB, SAP HANA,SQLSTATE, out of memory,EMFILE, read-only filesystem, a malformed database image, an oversized document) is recorded. The startup, process error and busy-write retry handlers record theirs too. - What WeKan measured. The database health probe
(
server/lib/databaseHealth.js, at startup and every five minutes on one server of the installation) records restarts, a data directory on a network filesystem, storage latency and disk space running out; startup records an index it had to create (server/lib/mongoStartup.js).
Each message is classified by the pure rules in models/lib/databaseErrors.js
(for example connection-lost, disk-full, auth-failed, mysql-syntax,
deadlock, timeout, duplicate-key, database-corrupt,
memory-exhausted, and unclassified for anything no rule knows). The same
problem on the same database is recorded at most once a minute.
A FerretDB canary - an operation WeKan never issues, such as server-side JavaScript, dropping a database or SQL that looks injected - is not a database problem: it goes to the Security Report instead.
The table
The same read-only event table as the other streams (Security Report), newest first, 10 rows a page, with a search box that filters as you type. On this stream:
| Column | Holds |
|---|---|
| Category | which database: mongodb, sqlite, postgresql, mysql, mariadb, hana or unknown |
| Name | the kind of problem, for example disk, auth, syntax, timeout or health |
| Severity | the rule's severity |
| Action | the classifier's rule id |
| Source | the operation, or Meteor._debug, or health |
| Detail | what it means and what to do, then — and what the database itself said |
The search also finds the database name, the rule id and words of the database's own message.
Acknowledge or clear the stream on Summary. Rows older than
EVENTLOG_RETENTION_DAYS (default 90) are deleted daily. The rows are also on
GET /api/admin/problems/database
(REST API).
Related
- Supported databases
- Recovery.md - SQLite corruption and the recovery history.
- FerretDB remediation design - the database canaries.