electric-db-collection: on-demand + persistence fails after relaunch with "Snapshot requests are not supported in full mode"
Los mantenedores suelen responder en 1 día
Evaluación
- Dificultad
- 3/5
- Tiempo estimado
- 1-2 días
- Aptitud para principiantes
- 68/100
- Tipo de issue
- Error
- Claridad
- Bien especificado
- Estado de actividad
- Activo
- Stack tecnológico
- react-native, sqlite, typescript
- Área
- database
Línea de trabajo
Start with loadSubset in dist/esm/electric.js around line 248, then trace how needsFullSnapshot selects full mode around line 828. Ask the reporter to share their two-launch node:test reproduction, which exercises electricCollectionOptions, persistedCollectionOptions, and createExpoSQLitePersistence. Done means a relaunch with tagged rows no longer requests a snapshot in full mode and the collection becomes ready.
Escrito por el modelo de indexación a partir del texto del issue.
Descripción
In short
An on-demand Electric collection with persistence works on the first app launch, but breaks on every launch after that. The second launch fails with:
Snapshot requests are not supported in full mode
After that error the collection stays in loading, and every live query on it fails until the local data is deleted.
It happens when the shape's rows have tags, which is the case when the shape's where has a subquery.
Versions
@tanstack/electric-db-collection0.5.4. The same code is in 0.5.5 and onmain.@electric-sql/client(the version 0.5.4 depends on)- Persistence:
persistedCollectionOptions+createExpoSQLitePersistence(React Native / Expo)
Setup
electricCollectionOptions({
syncMode: "on-demand",
shapeOptions: {
// the proxy adds a where with a subquery, for example:
// room_id IN (SELECT ... FROM permission_cache WHERE ...)
},
// ...
})
wrapped in persistedCollectionOptions(...).
Steps
- Launch the app. A live query loads a subset. Rows arrive with tags. The library saves a resume state with
requiresTagState: true. - Close the app, then open it again (a new JS process).
What goes wrong, step by step
Line numbers are from dist/esm/electric.js in 0.5.4.
- On the new launch, the tag state from the first launch is gone (it was in memory), so
retainsTagStateisfalse. The saved resume state saysrequiresTagState: true, soneedsFullSnapshotbecomestrue(~line 828). - Because of that, the stream is created in full mode:
logis not'changes_only', and the offset is-1(~line 872). - The live query then calls
loadSubset, andloadSubsetstill callsstream.requestSnapshot(...)(~line 248). - The Electric client does not allow snapshots in full mode, so it throws
Snapshot requests are not supported in full mode. - The throw comes before the stream starts, so the stream never starts and the collection never becomes ready.
So the library chooses full mode, then asks for something that full mode refuses.
Expected
When the stream is in full mode, loadSubset should not request a snapshot. It should wait until the full sync is ready, because the full log will contain the rows anyway.
Our workaround
We ship a small patch (about 35 lines) in loadSubset. When syncMode === 'on-demand' && needsFullSnapshot, it does not call requestSnapshot. It waits for the first markReady instead (it rejects if the initial sync calls markError, and resolves if the collection is aborted).
With the patch, the second launch works. One cost is left: every launch downloads the full shape again, because the tag state is not saved. Saving the tag state with the resume state may be the better fix. You know the design better than we do.
We are glad to open a PR with the patch, or share our test.
Test
We have a node:test test with two "launches". It uses the real electricCollectionOptions + persistedCollectionOptions + createExpoSQLitePersistence over node:sqlite, and a small fake Electric server that tags rows. It fails on 0.5.4 and passes with the patch. We can post it here.
Related
#811 has the same error text, in progressive mode.
- Lenguaje dominante
- TypeScript
- Estrellas
- 3.9k
- Forks
- 268
- Merge medio
- 1 d 2 h
- PR fusionados (30 d)
- 200
Preparar el entorno
- Sin Dockerfile ni archivo de Docker Compose
- Tiene una plantilla de pull request
- Leer la guía de contribución
Primeros pasos
- Lee el issue completo y luego la guía de contribución del proyecto.
- Comenta en el issue que vas a ocuparte — evita que dos personas hagan lo mismo.
- Haz un fork del repositorio y trabaja en una rama.
- Abre un pull request que haga referencia al número del issue.
Más de TanStack/db
-
Reusable queries: standalone descriptor resolution and nested alias composition limitsPosiblemente ocupada @KyleAMathews la tomó hoy. Abierto
Dificultad 5/5 Más de una semana Aptitud para principiantes 25/100
Los mantenedores suelen responder en 1 día
-
Dificultad 3/5 1-2 días Aptitud para principiantes 72/100
Los mantenedores suelen responder en 1 día
-
Dificultad 4/5 3-5 días Aptitud para principiantes 48/100
TanStack/db#1972 · 2 comentarios ·
Los mantenedores suelen responder en 1 día
-
Dificultad 5/5 Más de una semana Aptitud para principiantes 45/100
Los mantenedores suelen responder en 1 día
-
Dificultad 5/5 Más de una semana Aptitud para principiantes 45/100
Los mantenedores suelen responder en 1 día
Todos los issues de TanStack/db
Issues similares
-
[bug] diagnostics.dumpBody:Buffer 形态请求(透传 lane)跳过 dumps/ 落盘,仅留 raw/-unknown-Posiblemente ocupada @ranxianglei la tomó hoy. Abierto
Dificultad 2/5 1-3 horas Aptitud para principiantes 62/100
ranxianglei/billion-context#2421 · 2 comentarios ·
Los mantenedores suelen responder en 1 día
-
pending triage
Dificultad 2/5 1-3 horas Aptitud para principiantes 76/100
nuxt/test-utils#1842 ·
Los mantenedores suelen responder en 1 día
-
Dificultad 2/5 1-3 horas Aptitud para principiantes 85/100
MoonshotAI/kimi-code#4146 ·
Los mantenedores suelen responder en 1 día
-
Dificultad 2/5 1-3 horas Aptitud para principiantes 85/100
farbenmeer/tapi#531 ·
-
Dificultad 2/5 1-3 horas Aptitud para principiantes 85/100
Los mantenedores suelen responder en 1 día