usenet-indexer

Prober: a flush with a NUL in a name, or one article twice, no longer fails

Test and deploy #236:Commit d405b658c3 pushed by daan

rebuild

2026-10-18 17:50:46 +02:00

5m6s

Prober: no claims under PROBE_MIN_FREE_GB free on the NVMe

Test and deploy #235:Commit 9fb36d217c pushed by daan

rebuild

2026-10-18 17:42:41 +02:00

4m1s

probe_cache's primary key on the catalog's tablespace

Test and deploy #234:Commit de7c758206 pushed by daan

rebuild

2026-10-18 13:05:00 +02:00

4m7s

REBUILD.md: decision g, connections shared by priority

Test and deploy #233:Commit 38fd19b42e pushed by daan

rebuild

2026-10-18 11:33:47 +02:00

4m3s

NNTP shares: the account's connections shared by priority

Test and deploy #232:Commit 39480c486f pushed by daan

rebuild

2026-10-18 11:23:53 +02:00

3m39s

Prober: claim range per value of unreadable, from the index ends

Test and deploy #231:Commit 7eba53043a pushed by daan

rebuild

2026-10-18 11:05:05 +02:00

4m46s

The par2 prune's cursor is (created_at to the microsecond, set_id): 20k sets of one flush share a created_at

Test and deploy #230:Commit a5e0399ee6 pushed by daan

rebuild

2026-10-17 22:14:22 +02:00

3m57s

Fold only sidecar-only releases (sidecar_only at insert); no queue-wide scan for gone releases

Test and deploy #229:Commit f2c5607ec3 pushed by daan

rebuild

2026-10-17 21:01:43 +02:00

3m40s

The par2 prune gets 2 min per round (it held the merges back for hours); abandoned ranges past their end are finished

Test and deploy #228:Commit aa20d6b79f pushed by daan

rebuild

2026-10-17 20:47:56 +02:00

5m34s

Accounts with limits: a Newznab key and a tier per user, daily API and NZB counts, invites

Test and deploy #227:Commit 1322e42fb7 pushed by daan

monetization

2026-10-17 00:19:28 +02:00

2m33s

Poster search per poster found (LATERAL on releases_poster_newest_idx), at most 200 posters

Test and deploy #226:Commit c55918350e pushed by daan

rebuild

2026-10-16 18:03:45 +02:00

5m14s

Poster search through a posters table and releases_poster_newest_idx; complete releases are final; the tail before everything (-3)

Test and deploy #225:Commit c0e00f1707 pushed by daan

rebuild

2026-10-16 17:33:54 +02:00

8m18s

Postgres memory for the freed RAM: shared_buffers 8GB, effective_cache_size 32GB, random_page_cost 1.1, work_mem 16MB

Test and deploy #224:Commit c9a9c0155e pushed by daan

rebuild

2026-10-16 16:02:30 +02:00

5m15s

At most two leases per group: a group's batches wait for each other, more workers only queue

Test and deploy #223:Commit aa3bd82445 pushed by daan

rebuild

2026-10-16 15:35:11 +02:00

6m21s

The judge's unusable waits for final: next to an unread range a release's par2 or files may still come

Test and deploy #222:Commit b73e91219e pushed by daan

rebuild

2026-10-16 15:11:48 +02:00

5m28s

Settle every 5 min in rounds of 200k, not ~3M at once hourly; range blocks in a loop of their own

Test and deploy #221:Commit ec2e5a60be pushed by daan

rebuild

2026-10-16 15:03:25 +02:00

6m52s

Final completeness by time window: ranges keep their posting span, release_final() over blocks of done ranges

Test and deploy #220:Commit ce43099188 pushed by daan

rebuild

2026-10-16 14:30:19 +02:00

11m9s

Rules drop posters at the header pass: the CelebMovie family (rule set version 3)

Test and deploy #219:Commit 3d24db5565 pushed by daan

rebuild

2026-10-16 11:41:25 +02:00

4m35s

The judge hides scraped celebrity pictures; no ? in its patterns (a raw query takes it for a binding)

Test and deploy #218:Commit 093d0a71ad pushed by daan

rebuild

2026-10-16 11:06:18 +02:00

5m5s

A folder's images posted with their sizes are one release: no image set under the subject

Test and deploy #217:Commit 7d9a81784e pushed by daan

rebuild

2026-10-16 10:17:45 +02:00

4m19s

par2_files back on the NVMe: the merger reads it per set, a random read per row on the HDD

Test and deploy #216:Commit e154b2896f pushed by daan

rebuild

2026-10-16 08:27:25 +02:00

4m45s

Prune par2_files of complete and judged sets (hourly, from the merger)

Test and deploy #215:Commit 1a20fcd2c3 pushed by daan

rebuild

2026-10-16 08:06:13 +02:00

6m26s

Nightly pause (23:45) and resume (07:45); a deploy while paused leaves the workers stopped

Test and deploy #214:Commit dd35b49a3f pushed by daan

rebuild

2026-10-15 19:17:40 +02:00

4m12s

The sidecar fold every 5 min inside a long par2 pass too

Test and deploy #213:Commit 59c93071c7 pushed by daan

rebuild

2026-10-15 15:53:20 +02:00

4m47s

The sidecar fold first in a merger round: a queue, it needn't wait for the par2 pass

Test and deploy #212:Commit 2e25b7c56e pushed by daan

rebuild

2026-10-15 15:43:18 +02:00

5m18s

The merger keeps its place over a restart (worker_cursors), not from 1970 each deploy

Test and deploy #211:Commit e627862a8d pushed by daan

rebuild

2026-10-15 13:50:23 +02:00

4m53s

fragments back on the NVMe: the fragment merger reads every row, not by key

Test and deploy #210:Commit 46525f4595 pushed by daan

rebuild

2026-10-15 13:31:29 +02:00

5m14s

Fold sidecars (.nfo, .sfv, .nzb) apart from their upload into it

Test and deploy #209:Commit 329c7a1373 pushed by daan

rebuild

2026-10-15 12:15:50 +02:00

4m58s

The junk judge: releases judged by what their probes showed; verdicts hide them

Test and deploy #208:Commit 5897d595cf pushed by daan

rebuild

2026-10-15 11:51:29 +02:00

6m17s

par2_files, fragments and probe_cache with the chunks (CHUNK_TABLESPACE)

Test and deploy #207:Commit c65282ec5d pushed by daan

rebuild

2026-10-14 23:17:18 +02:00

5m34s