Hi Ufuk,

Here is the information regarding continuous job parameters. Please note
that I have not verified these with the latest version yet.

### 4 Key Parameters for Continuous Jobs
Continuous jobs never stop; they continuously cycle through a queue of
documents. These parameters control that cycle:

1. Recrawl Interval
  The minimum wait time before rechecking a document. Under adaptive
crawling, this interval automatically extends if no changes are found.
(Concept: "Do not recheck this document for at least X time.")
2. Maximum Recrawl Interval
  The maximum wait time allowed before forcing a recheck, capping the
adaptive backoff.
(Concept: "Force a recheck at least every X time, even if it never
changes.")
3. Expiration Interval
  The time a document can remain in the index without a successful recrawl.
If a page becomes unreachable or errors out for this duration, it is
removed.
(Concept: "Remove from index if unconfirmed after X time.")
4. Reseed Interval
  How often ManifoldCF looks for new URLs from the seed list. Without this,
the job only updates already-discovered pages.
(Concept: "Look for new pages every X time.")

### Scheduling a Continuous Job
Yes, setting a schedule makes sense, but it defines the allowed runtime
window (e.g., off-peak hours), not the crawl frequency. The actual recrawl
frequency is governed solely by the interval parameters above.

### Recommended Settings for Your Use Case
To reflect changes within 24 hours on a site that takes 2.5 days for a full
crawl, the following is an example configuration for your reference:

1. Recrawl Interval
  1–4 hours (quickly recheck active/changed pages).
2. Maximum Recrawl Interval
  24 hours (ensures every page is checked daily).
3. Expiration Interval
  48–72 hours (provides a buffer for transient network errors).
4. Reseed Interval
  24 hours (discovers newly added pages daily).

Regards,
Mingchun

2026年7月2日(木) 15:58 Ufuk YILMAZ via user <[email protected]>:

> Could someone explain me the meaning of the following 4 parameters for
> continuous jobs? I'm having a hard time understanding them from the
> documentation or through my trials.
>
> Recrawl interval (if continuous)
> Maximum recrawl interval (if continuous)
> Expiration interval (if continuous)
> Reseed interval (if continuous)
>
> Also, does setting a schedule for a continuous job make logical sense?
>  From my understanding, a continuous job continuously crawls a source,
> but on the UI there are fields that expect me to enter a
> month/week/day/hour etc.
>
> In my case, I'm crawling a website. If I start from zero (no index at
> all), the full crawl takes 2.5 days. However, not all pages change
> frequently, most of them stay the same for years. I'm aiming to reflect
> the changes the next day if possible.
>
> --Ufuk
>

Reply via email to