Commit Graph
15 Commits
Author SHA1 Message Date
jschoubben d6bd31a166 Say the try limits as what they are: a check when a try is asked
Tries let in while waiting may later all run at once; each node's build capacity bounds what runs
on it (review of issue 1890's try cap).
2026-10-26 20:11:08 +01:00
jschoubben c0a4b139a2 A try queued behind the build seat's checks no longer blocks the next try
Only tries a build node holds count toward the 4 at once; tries waiting for a node are bounded
apart at 8, and the refusal names both counts (issue 1890: at 19:53 four tries queued behind 44
checks for up to 47 minutes refused every further try).
2026-10-26 20:05:51 +01:00
jschoubben 6d60953f98 A running try names its asks and where each stands; build_usage says a seat's CPU-bound capacity
The checker wrote a try's asks only at its end, so a 31-minute try could not be told queued from
running (issue 1890). The try's record now names each ask as it is put, then the node that took
it, its state and the phase its step says, written on the step key's own changes. build_usage
names a running try's asks and takes a build module's capacity.<node> (configured bounded by its
node's build CPU, issue 1806's follow-up) when it is current with mesh-config.
2026-10-26 19:58:39 +01:00
jschoubben 60c7ffc981 Checker, builder and grants part: hear DEAD_LETTERS again until heard, say when they do not, and compose the same grant (issue 734, review round 1)
mesh-grants moves to the same mesh-sdk (3203358c), so its shadow user list grants the checker and the builder the
DEAD_LETTERS reads the controller's list will, and the take-over is not blocked by a difference. The checker and the
builder hear DEAD_LETTERS through buildrun.DeadLetterHearing: in the background, again at every reconnect and every ask
waited for while deaf, and keep whether they hear it in their new state hearing, which checker_status and builder_stuck
say as dead_letters. checker_try says that a try ended by the not-taken bound leaves its ask in the stream.

Not done as asked: a module health tool. The catalogue refuses a tool check on a module with no runtime of its own and no
other check beside it (ADR 0227 rule 8, ADR 0366 §1), so the deaf state is said in the status tools and the log.
2026-10-26 14:44:53 +01:00
jschoubben 6b85220fa4 Checker: an ask the build seat gives up on ends its try as error and asks a check again; a try no node takes ends at the step bound; checker_try_end ends one by hand (issue 734)
A check ask handed out five times during a bus reload was given up into DEAD_LETTERS and no build node wrote a step for it,
so try-1793017340003993788 stayed running for ever, held one of the four try places, and checker_try of its head kept
answering its id. The checker now hears the copies DEAD_LETTERS keeps of the build seats' given-up asks (an ordered
consumer from its start; the grant is mesh-sdk/mesh-controller DeadLetterReads, branch fix/734-lost-ask) and ends the
step it waits for: a try ends as error saying why, a check is asked again under its claim, up to three times, then fails
as not heard. A try whose ask no node takes within the step bound ends the same way (a check's backlog stays no fault,
ADR 0525 §9). checker_try_end ends a try by hand, freeing its place and stopping its judging.
2026-10-26 14:44:40 +01:00
jschoubben a7e51b416b checker_status reads the node-engine at its running commit, as the checker's decision does, so it stops listing every candidate as waiting for a trunk move (issue 568)
mesh/delivery merged: source.main moved to novox/mesh-delivery@ecf532af0e85
mesh/check being carried to a7e51b416bc1 on ecf532af0e85 | from 71f786f14662
2026-10-26 13:25:18 +01:00
jschoubben c66fe11e16 checker: claim no check of a repository an approved merge holds a turn on, and claim it when the turn goes (ADR 0548 §3; issue 568 D1) 2026-10-26 11:29:25 +01:00
jschoubben 9db15c7cf4 Builder and checker read tools read only the keys they need, in parallel, at most 1000 a call (issue 1827)
mesh/delivery merged: source.main moved to novox/mesh-delivery@b842b9422d2a
mesh/check checking 9db15c7cf4fc on b842b9422d2a
builder_stuck, checker_kept, checker_verdicts, checker_status, checker_previews, checker_pull and checker_try listed a
bucket of thousands of keys and read each value (builder_stuck also its history) one round trip at a time before
filtering, so they timed out at 30 s. Now keys are narrowed by what the key carries first, history is read only for rows
kept, values are read sixteen at a time, at most 1000 a call with the answer saying it was cut, steps newest first, a new
try reads only the newest 64 tries, and each tool logs the values read and the time taken. Same kind as the view's fix
(PR 52); the helper is copied, not put in mesh-sdk, as a few lines do not earn a group delivery.
2026-10-26 00:39:08 +01:00
jschoubben 69577f739d Judge hq as any repository, so its pull requests get a verdict and mesh/check from its Go tests and no script judges it (issue 1887 A10, ADR 0527 §2) 2026-10-23 16:11:43 +02:00
jschoubben d9d1e57e3a Delete the checker's hold, its own capacity and places, its node moves and its read of the controller's plans-building, so a check waits only in the build seats' streams (issue 1887 A5b, closes 1685; ADR 0525 §4, §7) 2026-10-23 16:09:34 +02:00
jschoubben 4ffe5f4d52 Start a never-judged head's check before the re-judges a trunk move asks, oldest first (issue 1897 S1)
mesh/delivery delivered
mesh/delivery-group group feat/fastcheck merged: every member is merged
mesh/check checking 4ffe5f4d5243 on 85aca46348a0 | novox/mesh-delivery#25 in group feat/fastcheck
A new pull request waited 25 to 45 minutes behind every open pull request's re-judge after a
merge. The semaphore becomes a ranked admission (ADR 0536 §4); checker_status names the
waiting checks by kind.
2026-10-23 14:09:42 +02:00
jschoubben e762028760 Make checker_status count only steps that can still run, so a key a build module left running is not read as a full node (issue 1890)
mesh/delivery delivered
mesh/check pass e76202876032 on b714eab995a5 | unit 422 run, 0 kept | integration 467 run, 0 kept | over bound: integration (slowest integration mesh-…
The checker's own places are released when a step ends or its wait passes the step bound; checker_status read every
'running' key in the build modules' steps buckets as a step in flight, so keys left by stopped, restarted or unassigned
build modules looked like full nodes. It now lists them under left_running with why (past the bound, module not on the
node, module not running), and leaves hq-only groups out of the unjudged candidates.
2026-10-23 09:03:45 +02:00
jschoubben 7abe230701 Judge a pushed head as a try on the checker, through checker_try and checker_try_read (ADR 0501 §5)
mesh-controller's try asked build-agent, which leaves the checking at R4e. A try is now a key in the checker's state
tries: the checker judges the head merged onto its trunk as it judges a pull request's head, through the same
judge, places and holds, writes its end in the key and owes the forge nothing (no status, no preview, no verdict).
One try per head, at most four at once; hq and a repository whose trunk the gate does not name are refused with
"no level of <repo> is checked by the checker".
2026-10-22 18:03:28 +02:00
jschoubben d83156fb21 Add checker_previews and checker_status read tools so the live acceptance of issues 1680 and 1685 can be read (C1, C2) 2026-10-22 17:25:11 +02:00
jschoubben 4cdba63b9b Read tools for the checker's verdicts, a pull request's report and the kept verdicts (issue 1686) 2026-10-21 21:32:08 +02:00