pg_cron's launcher connects to cron.database_name at postmaster start and is
registered BGW_NEVER_RESTART. On a host upgraded onto an existing /nsm/postgres
volume, init-db.sh never runs, so so_telegraf does not exist when PostgreSQL
starts -- the launcher dies and never retries. Salt then creates the database,
the extension, and the schedule, all of which succeed, but no worker is left to
fire the job. partman.run_maintenance_proc() therefore never runs: partitions
stop being premade after create_parent's initial window and every metric lands
in <parent>_default. Retention never fires either.
That state is self-perpetuating. Once the default partition holds rows for a day
with no child, PostgreSQL cannot create that child at all -- attaching it would
violate the default partition's constraint -- so maintenance aborts on the first
parent it reaches. Fixing the scheduler alone does not recover a stalled grid.
Point cron.database_name at the always-present postgres database and register the
job with cron.schedule_in_database targeting so_telegraf, so the launcher no
longer depends on database creation order. group_role drops any registration left
behind in so_telegraf, and both halves are guarded on the live GUC so applying
postgres.telegraf_users before the postgresql.conf change has restarted the
container skips instead of failing.
Maintenance now runs so_admin.telegraf_maintenance(), which drains stranded rows
out of any default partition before calling partman: expired rows are deleted,
the rest are repartitioned. It runs from the state on every highstate as well as
hourly from pg_cron, so a grid whose worker is dead still recovers on its own.
The routines live in a postgres-owned schema so_telegraf has no rights on, since
pg_cron executes them as postgres.
Existing grids are recovered by a marker-guarded repair state that truncates the
non-empty defaults once per host. The backlog is mostly past retention already
and moving tens of GB just to delete most of it is not worth the WAL.
Also raise premake from 3 to 7, reconciled onto existing parents in the retention
subcommand, so an outage has a week of headroom before anything reaches a
default partition, and add a check subcommand reporting partition age, default
occupancy and last job status.
init-db.sh only runs on a fresh PGDATA (docker-entrypoint-initdb.d), so a
cluster left partially initialized -- e.g. the container restarted mid first-init
by a watch trigger -- never recovered: the so_postgres role existed but its
schema grants and the so_telegraf database were missing, breaking SOC with
"permission denied for schema public".
- init-db.sh: make the role upsert race-safe. Try CREATE ROLE and fall back to
ALTER on duplicate_object/unique_violation instead of IF NOT EXISTS, so a
concurrent creator no longer aborts the script (set -e + ON_ERROR_STOP=1)
before the grants and so_telegraf creation run. The whole script is now
idempotent and safe to re-run.
- postgres/enabled.sls: move postgres_wait_ready here (was telegraf-gated in
telegraf_users.sls) and add postgres_bootstrap_soc_db, which re-runs init-db.sh
every highstate so a partially-initialized cluster self-heals.
- telegraf_users.sls: drop its now-duplicate postgres_wait_ready definition; it
resolves from postgres.enabled via the existing include.
- soup: remove bootstrap_so_soc_database and its post_to_3.2.0 call. The
highstates soup runs before postupgrade now reconcile the SOC DB, making the
one-shot bootstrap redundant.
The old flow had two writers for each per-minion Telegraf password
(so-minion wrote the minion pillar; postgres.auth regenerated any
missing aggregate entries). They drifted on first-boot and there was
no trigger to create DB roles when a new minion joined.
Split responsibilities:
- pillar/postgres/auth.sls (manager-scoped) keeps only the so_postgres
admin cred.
- pillar/telegraf/creds.sls (grid-wide) holds a {minion_id: {user,
pass}} map, shadowed per-install by the local-pillar copy.
- salt/manager/tools/sbin/so-telegraf-cred is the single writer:
flock, atomic YAML write, PyYAML safe_dump so passwords never
round-trip through so-yaml.py's type coercion. Idempotent add, quiet
remove.
- so-minion's add/remove hooks now shell out to so-telegraf-cred
instead of editing pillar files directly.
- postgres.telegraf_users iterates the new pillar key and CREATE/ALTERs
roles from it; telegraf.conf reads its own entry via grains.id.
- orch.deploy_newnode runs postgres.telegraf_users on the manager and
refreshes the new minion's pillar before the new node highstates,
so the DB role is in place the first time telegraf tries to connect.
- soup's post_to_3.1.0 backfills the creds pillar from accepted salt
keys (idempotent) and runs postgres.telegraf_users once to reconcile
the DB.
postgres_wait_ready requires docker_container: so-postgres, which is
declared in postgres.enabled. Running postgres.telegraf_users on its own
— as the reactor orch and the soup post-upgrade step both do — errored
because Salt couldn't resolve the require.
Include postgres.enabled from postgres.telegraf_users so the container
state is always in the render. postgres.enabled already includes
telegraf_users; Salt de-duplicates the circular include and the included
states are all idempotent, so repeated application is a no-op.
- firewall/map.jinja and postgres/telegraf_users.sls now pull the
telegraf output selector through TELEGRAFMERGED so the defaults.yaml
value (BOTH) is the source of truth and pillar overrides merge in
cleanly. pillar.get with a hardcoded fallback was brittle and would
disagree with defaults.yaml if the two ever diverged.
- Rename salt/postgres/files/pg_hba.conf.jinja to pg_hba.conf and drop
template: jinja from config.sls — the file has no jinja besides the
comment header.
The Telegraf backend selector lived at global.telegraf_output but it is
a Telegraf-scoped setting, not a cross-cutting grid global. Move both
the value and the UI annotation under the telegraf pillar so it shows
up alongside the other Telegraf tuning knobs in the Configuration UI.
- salt/telegraf/defaults.yaml: add telegraf.output: BOTH
- salt/telegraf/soc_telegraf.yaml: add telegraf.output annotation
- salt/global/defaults.yaml: remove global.telegraf_output
- salt/global/soc_global.yaml: remove global.telegraf_output annotation
- salt/vars/globals.map.jinja: drop telegraf_output from GLOBALS
- salt/firewall/map.jinja: read via pillar.get('telegraf:output')
- salt/postgres/telegraf_users.sls: read via pillar.get('telegraf:output')
- salt/telegraf/etc/telegraf.conf: read via TELEGRAFMERGED.output
- salt/postgres/tools/sbin/so-stats-show: update user-facing docs
No behavioral change — default stays BOTH.
- Deliver postgres super and app passwords via mounted 0600 secret files
(POSTGRES_PASSWORD_FILE, SO_POSTGRES_PASS_FILE) instead of plaintext env
vars visible in docker inspect output
- Mount a managed pg_hba.conf that only allows local trust and hostssl
scram-sha-256 so TCP clients cannot negotiate cleartext sessions
- Restrict postgres.key to 0400 and ensure owner/group 939
- Set umask 0077 on so-postgres-backup output
- Validate host values in so-stats-show against [A-Za-z0-9._-] before SQL
interpolation so a compromised minion cannot inject SQL via a tag value
- Coerce postgres:telegraf:retention_days to int before rendering into SQL
- Escape single quotes when rendering pillar values into postgresql.conf
- Own postgres tooling in /usr/sbin as root:root so a container escape
cannot rewrite admin scripts
- Gate ES migration TLS verification on esVerifyCert (default false,
matching the elastic module's existing pattern)
docker-entrypoint.sh runs the init-scripts phase with listen_addresses=''
(Unix socket only). The old pg_isready check passed there and then raced
the docker_temp_server_stop shutdown before the final postgres started.
pg_isready -h 127.0.0.1 only returns success once the real CMD binds
TCP, so downstream psql execs never land during the shutdown window.
pg_partman 5.x's create_partition() creates a per-parent template
table inside the partman schema at runtime, which requires CREATE on
that schema. Also extend ALTER DEFAULT PRIVILEGES so the runtime-
created template tables are accessible to so_telegraf.
Telegraf calls partman.create_parent() on first write of each metric,
which needs USAGE on the partman schema, EXECUTE on its functions and
procedures, and DML on partman.part_config.
- Telegraf's outputs.postgresql plugin uses Go text/template syntax,
not uppercase tokens. The {TABLE}/{COLUMNS}/{TABLELITERAL} strings
were passed through to Postgres literally, producing syntax errors
on every metric's first write. Switch to {{ .table }}, {{ .columns }},
and {{ .table|quoteLiteral }} so partitioned parents and the partman
create_parent() call succeed.
- Replace the \gexec "CREATE DATABASE ... WHERE NOT EXISTS" idiom in
both init-users.sh and telegraf_users.sls with an explicit shell
conditional. The prior idiom occasionally fired CREATE DATABASE even
when so_telegraf already existed, producing duplicate-key failures.
- Telegraf's partman template passed p_type:='native', which pg_partman
5.x (the version shipped by postgresql-17-partman on Debian) rejects.
Switched to 'range' so partman.create_parent() actually creates
partitions and Telegraf's INSERTs succeed.
- Added a postgres_wait_ready gate in telegraf_users.sls so psql execs
don't race the init-time restart that docker-entrypoint.sh performs.
- so-verify now ignores the literal "-v ON_ERROR_STOP=1" token in the
setup log. Dropped the matching entry from so-log-check, which scans
container stdout where that token never appears.
init-users.sh only runs on a fresh data dir, so upgrades onto an
existing /nsm/postgres volume never got so_telegraf. Pinning partman's
schema also makes partman.part_config reliably resolvable.
Every telegraf.* metric table is now a daily time-range partitioned
parent managed by pg_partman. Retention drops old partitions instead
of the row-by-row DELETE that so-telegraf-trim used to run nightly,
and dashboards will benefit from partition pruning at query time.
- Load pg_cron at server start via shared_preload_libraries and point
cron.database_name at so_telegraf so job metadata lives alongside
the metrics
- Telegraf create_templates override makes every new metric table a
PARTITION BY RANGE (time) parent registered with partman.create_parent
in one transaction (1 day interval, 3 premade)
- postgres_telegraf_group_role now also creates pg_partman and pg_cron
extensions and schedules hourly partman.run_maintenance_proc
- New retention reconcile state updates partman.part_config.retention
from postgres.telegraf.retention_days on every apply
- so_telegraf_trim cron is now unconditionally absent; script stays on
disk as a manual fallback
Per-minion schemas cause table count to explode (N minions * M metrics)
and the per-minion revocation story isn't worth it when retention is
short. Move all minions to a shared 'telegraf' schema while keeping
per-minion login credentials for audit.
- New so_telegraf NOLOGIN group role owns the telegraf schema; each
per-minion role is a member and inherits insert/select via role
inheritance
- Telegraf connection string uses options='-c role=so_telegraf' so
tables auto-created on first write belong to the group role
- so-telegraf-trim walks the flat telegraf.* table set instead of
per-minion schemas
- so-stats-show filters by host tag; CLI arg is now the hostname as
tagged by Telegraf rather than a sanitized schema suffix
- Also renames so-show-stats -> so-stats-show
Introduces global.telegraf_output (INFLUXDB|POSTGRES|BOTH, default BOTH)
so Telegraf can write metrics to Postgres alongside or instead of
InfluxDB. Each minion authenticates with its own so_telegraf_<minion>
role and writes to a matching schema inside a shared so_telegraf
database, keeping blast radius per-credential to that minion's data.
- Per-minion credentials auto-generated and persisted in postgres/auth.sls
- postgres/telegraf_users.sls reconciles roles/schemas on every apply
- Firewall opens 5432 only to minion hostgroups when Postgres output is active
- Reactor on salt/auth + orch/telegraf_postgres_sync.sls provision new
minions automatically on key accept
- soup post_to_3.1.0 backfills users for existing minions on upgrade
- so-show-stats prints latest CPU/mem/disk/load per minion for sanity checks
- so-telegraf-trim + nightly cron prune rows older than
postgres.telegraf.retention_days (default 14)
Mike Reeves
Josh Patterson