When to reload Asterisk and when it makes things worse

Reloading Asterisk feels like the obvious move after a config change, the same way you would restart a service to make a setting take. On a live dialer it is not that simple. A reload re-reads configuration, and depending on what you reload it can interrupt active calls. Most of the time on VICIdial you do not need to touch it at all.

What a reload actually does

Asterisk is the telephony engine underneath VICIdial. A reload tells it to read its configuration files again and apply the changes without a full restart. From the Asterisk command line you might run a broad core reload, or a narrow one like dialplan reload that only re-reads the Dialplan. A restart is heavier still: it tears down the whole process and brings it back up, which always drops every live Channel.

The narrower the reload, the safer it is. Reloading just the dialplan rarely disturbs calls already in progress. A full module reload can re-initialise channel drivers, and that is where active calls can get dropped mid-conversation.

Why you usually do not reload by hand

VICIdial manages most of its own Asterisk configuration. When you change a carrier, a phone, a server setting, or a campaign detail in the admin screen, VICIdial flags that change with rebuild_conf_files set to Y. The Keepalive process, which runs about once a minute, sees that flag, regenerates the affected Conf file, and reloads only what is needed. You make the change in the web interface and wait a minute, and it applies itself.

When a reload genuinely helps

• You edited an Asterisk config file directly, outside VICIdial's control, and need it to take effect. VICIdial will not rebuild a file it does not manage, so a manual reload is the only way to apply it.
• You changed something the keepalive does not cover, and you have confirmed the rebuild cycle will not pick it up on its own.
• You are debugging and need a known-clean state to test against, during a maintenance window with no agents dialing.

When it makes things worse

flowchart TD
  A[You changed something] --> B{Changed in VICIdial admin?}
  B -->|Yes| C[Set rebuild_conf_files Y already]
  C --> D[Keepalive rebuilds and reloads in about a minute]
  D --> E[Do nothing, just wait]
  B -->|No, edited a file directly| F{Are agents live on calls?}
  F -->|Yes| G[Wait for a quiet window]
  F -->|No| H[Reload now, narrow scope first]
  G --> H

A broad reload during a live shift is the classic way to make a small problem big. Calls in progress can drop, the predictive engine can stumble, and agents see their Agent session hiccup right as the dialer is at full pace. If your Drop rate suddenly spikes after someone ran a reload mid-shift, that is not a coincidence. The cost of waiting for a quiet window is almost always lower than the cost of dropped live calls.

The safe default

Make your change in the VICIdial admin, wait a minute for the keepalive to rebuild the conf files, and only hand-reload if you edited something outside VICIdial's control — and even then, in a quiet window. If after all that calls still are not going through, the problem is probably elsewhere; start with fixing calls that will not connect and the broader VICIdial troubleshooting playbook.

Knowing when to leave Asterisk alone is half the job of running a dialer. VICIfast hands you a managed, hardened VICIdial box where the keepalive and conf rebuilds are already wired up correctly, live in under 40 seconds. See our plans and pricing.