ephpm cache

Manage the OPcache from the CLI. Currently only the reset subcommand is implemented; status is planned. See the OPcache clustering roadmap.

Synopsis

ephpm cache [--host HOST] [--port PORT] <subcommand>

Subcommands

reset [--site NAME | --all]

Invalidate the OPcache for one vhost (or every vhost via the broadcast key). Functionally identical to ephpm deploy — both write the same opcache:version:<vhost> key via the RESP listener. The separate command exists so operators can distinguish a local dev reset from a deploy event in shell history or audit logs. On a cluster, both propagate via gossip because the RESP write lands in the same in-process KV.

# Reset a single vhost
ephpm cache reset --site blog

# Reset every vhost (broadcast)
ephpm cache reset --all

# Single-node / no sites_dir
ephpm cache reset

The running server must have [kv.redis_compat] enabled = true — the CLI is a separate process from the server, so it cannot poke the in-process KV DashMap directly. If the RESP listener is not reachable the CLI prints a hint pointing at the config knob.

status — planned, not yet implemented

Global and per-vhost OPcache stats (hit rate, script count, memory). The design is in the roadmap but the subcommand is not shipped. Use opcache_get_status() from PHP for now.

See also

• ephpm deploy — the deploy-shaped variant of reset
• OPcache clustering roadmap
• [opcache] config