K-S6 (#429). Every failure to open the data store at start-up used to exit with 1 and the driver's message, so nothing could tell a database that's down from a wrong password. The chart's pre-flight check (K-C5) needs to.
Each case now gets its own first words, an exit code from sysexits.h, and a one-line hint: bad configuration 78, unreachable 69, credentials refused 77, TLS failed 76, anything else 1. PostgreSQL errors are told apart fully. Other stores report bad configuration and otherwise 1. The table is in crates/store/src/registry/startup.rs.
I ran the release build against nine broken configs on the lab's PostgreSQL: bad JSON, unknown store type, missing secret file, refused, unknown host, timeout, wrong password, missing database, and TLS the server doesn't offer. Each one gave the expected code and line. The lab's five server pods run this build and pass the sign-in test. Unit tests cover the classification.
Docs: inbuxa/inbuxa.org PR, held for rc.32.
K-S6 (#429). Every failure to open the data store at start-up used to exit with 1 and the driver's message, so nothing could tell a database that's down from a wrong password. The chart's pre-flight check (K-C5) needs to.
Each case now gets its own first words, an exit code from sysexits.h, and a one-line hint: bad configuration 78, unreachable 69, credentials refused 77, TLS failed 76, anything else 1. PostgreSQL errors are told apart fully. Other stores report bad configuration and otherwise 1. The table is in `crates/store/src/registry/startup.rs`.
I ran the release build against nine broken configs on the lab's PostgreSQL: bad JSON, unknown store type, missing secret file, refused, unknown host, timeout, wrong password, missing database, and TLS the server doesn't offer. Each one gave the expected code and line. The lab's five server pods run this build and pass the sign-in test. Unit tests cover the classification.
Docs: inbuxa/inbuxa.org PR, held for rc.32.
Every start-up failure to open the data store exited with 1 and the
driver's own words, so a supervisor, a Kubernetes pre-flight check or a
person reading pod status couldn't tell a database that's down from a
wrong password. Now each case has its own first words and exit code
(sysexits.h), and a one-line hint:
78 data store: bad configuration settings file, secret files, database name
69 data store: unreachable refused, unknown host, timed out, starting, full
77 data store: credentials refused wrong user or password, missing permission
76 data store: TLS failed handshake failed
1 data store: error anything else
PostgreSQL errors are told apart fully; other stores report bad
configuration and otherwise 1. K-S6.
Blocking a user prevents them from interacting with repositories, such as opening or commenting on pull requests or issues. Learn more about blocking a user.
K-S6 (#429). Every failure to open the data store at start-up used to exit with 1 and the driver's message, so nothing could tell a database that's down from a wrong password. The chart's pre-flight check (K-C5) needs to.
Each case now gets its own first words, an exit code from sysexits.h, and a one-line hint: bad configuration 78, unreachable 69, credentials refused 77, TLS failed 76, anything else 1. PostgreSQL errors are told apart fully. Other stores report bad configuration and otherwise 1. The table is in
crates/store/src/registry/startup.rs.I ran the release build against nine broken configs on the lab's PostgreSQL: bad JSON, unknown store type, missing secret file, refused, unknown host, timeout, wrong password, missing database, and TLS the server doesn't offer. Each one gave the expected code and line. The lab's five server pods run this build and pass the sign-in test. Unit tests cover the classification.
Docs: inbuxa/inbuxa.org PR, held for rc.32.