A cache storage that is not available is logged and skipped; Souin then serves from its unbounded in-memory storage
Dieses Issue hat noch niemand übernommen.
Bewertung
- Schwierigkeit
- 4/5
- Geschätzter Aufwand
- 3-5 Tage
- Anfängerfreundlichkeit
- 48/100
Rechercherichtung
Start in dispatch.go at parseStorages and trace how configured storage UUIDs are registered and looked up during Provision, including the call immediately after s.parseStorages(ctx). Reproduce the large size lookup mismatch and failed SimpleFS provisioning described in the issue. Done means provisioning fails when a configured storage is unavailable, so Caddy does not silently serve from the in-memory default.
Vom Indexierungsmodell aus dem Issue-Text verfasst.
Beschreibung
Versions: Caddy v2.11.4, cache-handler v0.17.0, github.com/darkweak/storages/simplefs/caddy v0.0.20 (built with xcaddy, nothing else).
When the storage a cache block configures is not available, cache-handler logs it and Souin serves every response from its in-memory default storage. Caddy accepts the configuration, caddy reload included, so its automatic rollback never applies, and every cached response then lives in Caddy's heap with no cap, although the configuration asked for a disk-backed, size-capped storage.
Two ways to get there, both reproduced on the versions above with this Caddyfile (only the configuration block changes):
{
admin off
cache {
ttl 60s
simplefs {
configuration {
path /tmp/simplefs
size 1000000
}
}
}
}
:8080 {
cache
header Cache-Control "public, max-age=60"
respond "hello" 200
}
1. size of 1,000,000 or more. The storer provisions and creates its directory, but the lookup does not find it: the name cache-handler composes for the lookup formats the decoded size (a float64) as 1e+06, while the storer registers under 1000000. The only log line is Souin's warning:
{"level":"warn","logger":"http.handlers.cache","msg":"You're running Souin with the default storage that is not optimized and for development purpose. ..."}
and every hit reports Cache-Status: Souin; hit; ttl=59; key=GET-http-127.0.0.1:8080-/; detail=DEFAULT, with nothing written under /tmp/simplefs. With size 1000 the same Caddyfile serves detail=SIMPLEFS and writes the body files.
2. A storage that fails to provision, e.g. path /nope/simplefs (a directory the process cannot create). parseStorages (dispatch.go) logs the error and carries on:
{"level":"error","logger":"http.handlers.cache","msg":"Error during SimpleFS init, did you include the SimpleFS storage (--with github.com/darkweak/storages/simplefs/caddy)? provision storages.cache.simplefs: mkdir /nope: read-only file system"}
then the same warning, and the same detail=DEFAULT hits.
Suggested fix: fail Provision when a configured storage (Found) is not registered under the UUID Souin looks it up by. That covers both cases: a reload is then refused and Caddy keeps the config it ran, and a start fails. We run this as a small patch, called right after s.parseStorages(ctx) and before middleware.NewHTTPCacheHandler:
func (s *SouinCaddyMiddleware) requireConfiguredStorages() error {
dc := s.Configuration.DefaultCache
for _, configured := range []struct {
name string
provider configurationtypes.CacheProvider
}{
{"badger", dc.Badger}, {"etcd", dc.Etcd}, {"nats", dc.Nats}, {"nuts", dc.Nuts},
{"olric", dc.Olric}, {"otter", dc.Otter}, {"redis", dc.Redis}, {"simplefs", dc.SimpleFS},
} {
if configured.provider.Found &&
(configured.provider.Uuid == "" || core.GetRegisteredStorer(configured.provider.Uuid) == nil) {
return fmt.Errorf("the %s storage is configured but is not available (its error is logged above): refusing to serve from the in-memory storage", configured.name)
}
}
return nil
}
Case 1 could also be fixed at the source, by formatting size as an integer in the lookup name, or by asking the provisioned storer for the name it registered under. Happy to open a PR for either.
- Vorherrschende Sprache
- Go
- Sterne
- 395
- Forks
- 30
- Ø Merge
- 32 Min.
- Gemergte PRs (30 T.)
- 1
Entwicklungsumgebung
Dieses Projekt bietet weder Dev-Container noch Dockerfile noch Beitragsleitfaden – die Einrichtung liegt bei Ihnen. Beginnen Sie mit der README; die allgemeinen Schritte stehen in unserem Leitfaden für den ersten Beitrag.
Erste Schritte
- Lesen Sie das ganze Issue und danach den Beitragsleitfaden des Projekts.
- Schreiben Sie ins Issue, dass Sie es übernehmen — das erspart doppelte Arbeit.
- Forken Sie das Repository und arbeiten Sie in einem Branch.
- Öffnen Sie einen Pull Request, der die Issue-Nummer nennt.
Mehr aus caddyserver/cache-handler
-
Schwierigkeit 1/5 Unter einer Stunde Anfängerfreundlichkeit 72/100
caddyserver/cache-handler#138 · 1 Kommentar · 3 Reaktionen ·
-
Schwierigkeit 3/5 Ein halber Tag Anfängerfreundlichkeit 45/100
caddyserver/cache-handler#146 ·
-
Admin API purge/invalidation still panics with nil pointer on v0.16.0 (same root cause as #140)Offen
Schwierigkeit 3/5 1-2 Tage Anfängerfreundlichkeit 76/100
caddyserver/cache-handler#143 · 2 Kommentare ·
-
Can not flush cacheOffen
Schwierigkeit 3/5 1-2 Tage Anfängerfreundlichkeit 55/100
caddyserver/cache-handler#140 · 4 Kommentare · 1 Reaktion ·
-
`Incomming` has a typoOffen
Schwierigkeit 1/5 Unter einer Stunde Anfängerfreundlichkeit 45/100
caddyserver/cache-handler#139 · 1 Kommentar ·
Alle Issues in caddyserver/cache-handler
Ähnliche Issues
-
Schwierigkeit 2/5 1-3 Stunden Anfängerfreundlichkeit 81/100
apache/rocketmq-dashboard#6895 ·
Maintainer antworten meist innerhalb von 3 Tagen
-
Schwierigkeit 1/5 Unter einer Stunde Anfängerfreundlichkeit 87/100
777genius/agent-notifications#490 ·
Maintainer antworten meist innerhalb von 1 Tag
-
bug
Schwierigkeit 2/5 1-3 Stunden Anfängerfreundlichkeit 78/100
devantler-tech/ksail#7836 ·
Maintainer antworten meist innerhalb von 1 Tag
-
Schwierigkeit 2/5 1-3 Stunden Anfängerfreundlichkeit 76/100
keyxmakerx/Chronicle#1469 ·
Maintainer antworten meist innerhalb von 1 Tag
-
Schwierigkeit 2/5 1-3 Stunden Anfängerfreundlichkeit 73/100
Maintainer antworten meist innerhalb von 1 Tag