set a limit for how far back we'll construct new runs of existing jobs for new hosts

2 files changed, 11 insertions, 3 deletions

diff --git a/src/dbctx.rs b/src/dbctx.rs
index f5f900a..e224310 100644
--- a/src/dbctx.rs
+++ b/src/dbctx.rs
@@ -549,10 +549,18 @@ impl DbCtx {
     }
 
     pub fn jobs_needing_task_runs_for_host(&self, host_id: u64) -> Result<Vec<Job>, String> {
+        // for jobs that this host has not run, we'll arbitrarily say that we won't generate new
+        // runs for jobs more than a day old.
+        //
+        // we don't want to rebuild the entire history every time we see a new host by default; if
+        // you really want to rebuild all of history on a new host, use `ci_ctl` to prepare the
+        // runs.
+        let cutoff = crate::io::now_ms() - 24 * 3600 * 1000;
+
         let conn = self.conn.lock().unwrap();
 
         let mut jobs_needing_task_runs = conn.prepare(sql::JOBS_NEEDING_HOST_RUN).unwrap();
-        let mut job_rows = jobs_needing_task_runs.query([host_id]).unwrap();
+        let mut job_rows = jobs_needing_task_runs.query([cutoff, host_id]).unwrap();
         let mut jobs = Vec::new();
 
         while let Some(row) = job_rows.next().unwrap() {
diff --git a/src/sql.rs b/src/sql.rs
index e98c111..1071279 100644
--- a/src/sql.rs
+++ b/src/sql.rs
@@ -144,9 +144,9 @@ pub const PENDING_RUNS: &'static str = "\
 
 pub const JOBS_NEEDING_HOST_RUN: &'static str = "\
     select jobs.id, jobs.source, jobs.created_time, jobs.remote_id, jobs.commit_id, jobs.run_preferences from jobs \
-    where jobs.run_preferences=\"all\" \
+    where jobs.run_preferences=\"all\" and jobs.created_time > ?1 \
     and not exists \
-        (select 1 from runs r2 where r2.job_id = jobs.id and r2.host_id = ?1);";
+        (select 1 from runs r2 where r2.job_id = jobs.id and r2.host_id = ?2);";
 
 pub const ACTIVE_RUNS: &'static str = "\
     select id,