Replaces the bare signal(SIGINT, sigbreak) install with sigaction() and
SA_RESTART cleared, so a blocking read reliably returns EINTR instead of
being silently restarted by glibc's BSD signal() semantics. Adds
xcetest_interrupted() as a read-only accessor over the existing break_out
flag for the interruptible-call adapter (tasks 1.4/1.5) to consume.
Proven by a real fork+SIGINT+blocked-read test that first demonstrates
the old bug, then the fix. 1280/1280 controller suite passing (+2).
Task 1.3 of .kiro/specs/xce-test-cli-ux/tasks.md.
Requirements: 9.9
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_019jaMf9QEwpugH2W9KAuc7f
flagspec_reject() now names a likely-intended flag when an unrecognized
flag closely matches exactly one entry in the descriptor table (genuine
Levenshtein DP, silent on zero or ambiguous matches). flagspec_render_help()
generates the short/long/required-optional/default flag-listing block from
a descriptor table. 1278/1278 controller suite passing (+5).
Task 1.2 of .kiro/specs/xce-test-cli-ux/tasks.md.
Requirements: 6.1, 6.2, 6.3, 6.4, 8.2, 10.1, 10.2, 10.3
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_019jaMf9QEwpugH2W9KAuc7f
Adds flagspec_t descriptor table, getopt_long option-table builder, and
the shared stderr-routed rejection path that every flag-accepting
command will migrate onto in later tasks. TDD'd via test_flagspec.c
(1273/1273 controller suite passing).
Task 1.1 of .kiro/specs/xce-test-cli-ux/tasks.md.
Requirements: 2.1, 2.2, 2.3, 2.4, 3.1
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_019jaMf9QEwpugH2W9KAuc7f
The previous fix for the "slots" regression (5b18f14a) worked but was an
unneeded special case: strcmp(vres->res->name, "slots") to fall back to
nslots directly. alloc_req->host already holds exactly that value during
the global phase -- both allocators' own unconditional "add slots" step
stages it there regardless of phase, and plugin_prepare_allocation()
never resets .host during the global phase, so by the time sched_fill_
limit_requests() runs (right after, same invocation), .host contains
exactly "slots"'s job-wide total and nothing else.
Read alloc_req->host FIRST for this combination and fall back to the
rusages-based on-demand computation only when nothing is staged there
(get_resval_by_id()'s "not found" == 0.0, safe here since "slots" is
never legitimately 0 -- seqjob only runs for maxTasks==1, parjob only
for maxTasks>1). Data-driven instead of name-based, matching bug-1789's
own earlier flag-based-dispatch philosophy, and generalizes for free to
any future resource an allocator plugin similarly pre-stages outside
its normal rusages loop.
test_fill_limit_requests_global_limit_on_slots_reads_host now pre-stages
alloc_req->host exactly as the allocators do, rusages left empty,
matching production exactly.
Full xce-build.sh all asan --incremental: 94/94 Ok. Live re-verified
against both the original gitea#46 scenario and the ug1_global_mem
scenario: byte-identical scheduler traces to every prior verification
round.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GLfB5uWRdmx2WiZSFC7Fm7
The limit-plugin-ownership refactor (fa34b038) broke gitea#46's exact
original scenario. calc_wanted_job_total_resval() scans the job's
rusages by name, but "slots" is never a declared rusage entry --
seqjob_prepare_allocation()/parjob_prepare_allocation() derive its want
directly from the job's own task count via their own unconditional
key_find("slots", ...) + store, entirely outside the rusages loop. The
refactor's on-demand lookup therefore always missed for "slots",
silently leaving the computed want at 0 -- reintroducing bug-1789's
exact original symptom (matching worked, but caps never bound).
Live-confirmed against a fresh deploy of gitea#46's own repro (9 plain
sleep jobs, user1, queue normal, no -R flag): limit1/ug1_slots/
ug1_normal_slots all matched correctly but zero unsatisfy_ lines
appeared and all 9 jobs dispatched uncapped. The refactor's own new
unit test had masked this -- it artificially added a "slots" rusages
entry that doesn't reflect production reality.
Fix: sched_fill_limit_requests() now falls back to `nslots` directly
(the same value both allocators already store as "slots"'s own
per-host want) whenever the rusages lookup misses and the resource is
named "slots" -- the one documented exception to the rusages-lookup
rule, matching this codebase's existing "slots" special-casing
convention.
New test test_fill_limit_requests_global_limit_on_slots_uses_nslots
proves the fallback with no "slots" entry in rusages at all. The
original test reworked to use "mem" instead, so it now genuinely
exercises the general on-demand path.
Full xce-build.sh all asan --incremental: 94/94 Ok. Live re-verified
against both the original gitea#46 scenario (byte-identical to the
pre-refactor trace) and the ug1_global_mem scenario -- both correct.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GLfB5uWRdmx2WiZSFC7Fm7
bug-1790's fix worked but put limit-plugin-specific logic in the wrong
place: seqjob_prepare_allocation()/parjob_prepare_allocation() (the
allocator plugins) unconditionally pre-computed a job-wide total for
every per-host resource in the job's rusages, "just in case" some
global limit needed one. That's the allocator plugins' concern leaking
into limit evaluation -- duplicated across both allocators (a third
would need it too), and wasteful (computed for resources no limit
actually references).
Moved ownership to sched_fill_limit_requests() (resource.c), called
once centrally from plugin_prepare_allocation() regardless of which
allocator ran:
- Removed calc_wanted_host_resval_job_totals() and its calls from both
allocator plugins entirely.
- Added calc_wanted_job_total_resval() (resreq.c): an on-demand,
single-resource lookup by name (no resmgr/key_find needed -- the
caller already has the resolved xce_resource_t), computing the same
job-wide-total formula only for the one (scope, underlying-resource)
combination that needs it -- a GLOBAL limit on a per-host resource.
The other three combinations are unchanged, still reading straight
out of alloc_req.
- sched_fill_limit_requests() gains rusages/nhosts/nslots parameters
back (a genuine need this time, not a re-added unused one).
plugin.c's two call sites pass obj->u.primary->resreq->rusages
(NULL for COMPOUND/ALTER, matching this codebase's existing scope
convention for those tags) and ctx->nhosts/ctx->nslots.
All 23 testsuite link-satisfying stubs mechanically updated. Removed
the two allocator-level regression tests from the first pass (the
behavior they proved no longer exists by design); rewrote
test_fill_limit_requests_global_limit_on_host_resource_reads_host to
exercise the new on-demand path directly, nslots=3 to prove the job's
WHOLE total is used.
Full xce-build.sh all asan --incremental: 94/94 Ok, zero regressions.
Live re-verified against the same ug1_global_mem deploy/scenario:
byte-identical scheduler trace to the pre-refactor result, confirming
behavior-preserving.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GLfB5uWRdmx2WiZSFC7Fm7
test_job_rusage_reconciles_mem_reservation_against_usage asserted
w->resInUse was already charged 4.0 for a JOBS_PUSH-dispatched RUN job,
immediately after the push with no round boundary simulated. That
assumption was wrong, not the scheduler: resInUse/sharedInUse is never
charged incrementally at JOBS_PUSH/dispatch time. It's fully re-derived
once per round by sched_account_jobs() (scheduler.c), which iterates
every currently-RUN job and calls sched_charge_account() on each one's
accounts -- called by both the real normal-mode driver (driver.c) and
the sim driver (sim_loop.c), always before anything else in that round
reads resInUse. Confirmed by reading the full JOBS_PUSH handler
(core_events.c) and builtin_fcfs_on_job_push(): neither ever charges or
uncharges resInUse, by design -- core_session.c's own doc comment on
core_begin_session_internal() already documents this "recompute from
source of truth each round" architecture.
Fix: the test now calls sched_charge_account(&core, tracked->accounts)
right after JOBS_PUSH, simulating the one round-boundary reconciliation
step sched_account_jobs() would perform for its single RUN job (that
function itself lives in scheduler.c, not white-box #included here, so
its single-job effect is reproduced directly via the same primitive it
calls internally). No production code changed.
Full scheduler suite: 77/77 Ok (was 76/77). Full xce-build.sh all asan
--incremental: 94/94 Ok, zero regressions.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GLfB5uWRdmx2WiZSFC7Fm7
free_host_account() (src/lib/misc.c) freed acct->accounts[] and acct
itself but never acct->host, the strdup'd string make_host_account()
(xcealloc.c) allocates -- a real leak on every free_job_account() call,
including the live production dispatch path (scheduler.c, core_events.c),
not just test fixtures. Already root-caused and the exact fix specified
in bug-1621 (logged, never applied); applied here: FREEUP(acct->host)
before FREEUP(acct), mirroring free_res_account()'s existing
FREEUP(acct->resname) pattern.
Also closes bug-1723: test_xcealloc.c's two multi-host overflow
regression tests (added by bug-1722's fix) never freed their own
locally-built fixture trees. Added free_JobAlloc_ptr(&jalloc) to
test_make_job_account_multi_host_no_overflow() and
free_HostAlloc_list(2, hallocs) to
test_make_host_allocs_multi_host_no_overflow(), exactly as bug-1723's
own root-cause note specified.
Together these take xcelib_test from a deterministic 456-byte/29-
allocation LeakSanitizer SIGABRT (masking every other libxce test result
behind it) to 1/1 Ok. Full xce-build.sh all asan --incremental: 94/94 Ok,
zero regressions.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GLfB5uWRdmx2WiZSFC7Fm7
Adds ug1_global_mem to limits.yaml: a permanent, real GLOBAL limit on
"mem" (a per-host resource, not "slots"), capping ug1's total
cluster-wide mem usage at 5GB. This is exactly the scenario bug-1790
fixed but only unit-tested until now -- no limit in the shipped example
config exercised a global cap on a non-slots per-host resource.
Live-verified via a fresh single-node deploy: 3 jobs submitted as user1
(ug1 member) each requesting rusage[mem=2000000000] (2GB) on queue
normal. Scheduler trace confirms ug1_global_mem correctly matches and,
for the first time, computes a real nonzero job-wide-total wanted value:
unsatisfy_ resource 36 total 5000000000.000000 inuse 4000000000.000000
uncomit 0.000000 wanted 2000000000.000000
unsatisfy_ unmet resource 36 (limit=ug1_global_mem) shortfall
1000000000.000000
The 3rd job (4GB inuse + 2GB wanted = 6GB > 5GB cap) correctly PENDed
until an earlier job's mem freed up. Deploy config validated with zero
new warnings.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GLfB5uWRdmx2WiZSFC7Fm7
Closes the gap bug-1789 explicitly carved out: a GLOBAL limit built on a
per-host resource other than "slots" (e.g. a future absolute global mem
cap) needs ctx->alloc_req.host to carry that resource's job-wide total
want during the global phase, but only "slots" ever got that treatment
(via seqjob/parjob_prepare_allocation()'s own unconditional store).
Adds calc_wanted_host_resval_job_totals() (resreq.c), sharing its loop
body with calc_wanted_global_resvals() via a new static helper
parameterized only by which XCE_RESF_GLOBAL direction to match -- it
computes the SAME job-wide-total formulas (PER_TASK: value*nslots, PER_JOB:
value, PER_HOST: value*nhosts) for non-global resources instead of global
ones. seqjob/parjob_prepare_allocation()'s global-phase branch now calls
it into ctx->alloc_req.host alongside the existing calc_wanted_global_
resvals() call into .shared.
Safety was traced, not assumed: plugin_prepare_allocation() only resets
.host at the top of a per-host call, never during the global phase, and
sched_fill_limit_requests() for glimits reads .host later in the SAME
global-phase invocation -- transferring the value into the limit's
permanent .shared record before any later per-host call ever resets
.host. An earlier candidate fix (reusing calc_wanted_host_resvals()
directly) was rejected during design: its PER_HOST formula and, for
parjob, its per-bucket nslots don't give a job-wide total.
New regression tests in test_seqjob_allocator.c/test_parjob_
combinatorial.c prove a non-"slots" per-host resource ("mem") now gets
correctly populated; parjob's test uses maxTasks=3 specifically to prove
the formula uses the job's whole nslots, not a bucket's share.
Full suite (xce-build.sh all asan --incremental): 92/94, both failures
pre-existing and unrelated (bug-1762, bug-1733's already-documented
xcealloc.c leak). Not live-verified -- no limits.yaml entry exercises
this path yet.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GLfB5uWRdmx2WiZSFC7Fm7
meson/ninja/capnp/Go only exist inside the hermetic nix dev shell, but
xce-build.sh never entered it itself -- the /xce-build skill documented
wrapping every invocation in `nix develop ./src --command ...` by hand,
so a bare `bash src/tools/xce-build.sh ...` (run directly instead of
through the skill) failed opaquely with "meson: command not found". This
was a recurring, previously-logged gap that resurfaced again today mid
bug-1789 verification.
The script now checks IN_NIX_SHELL at the top and re-execs itself under
`nix develop` if unset, so every caller -- the skill, a bare invocation,
CI, a human at a terminal -- gets it automatically. A nested call (already
inside the shell) is a no-op past that check. README.md's daily-develop
section updated to match; the old "or use nix develop --command" example
is now redundant and removed.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GLfB5uWRdmx2WiZSFC7Fm7
sched_fill_limit_requests() used to read a limit's underlying resource's
want from the same vector its own virtual-resource entry gets written to.
That's wrong whenever the two differ in scope -- e.g. a GLOBAL limit built
on top of a per-host resource like "slots" (limits.yaml's
ug1_normal_slots): the limit's own vres entry belongs in alloc_req->shared,
but "slots"'s want is only ever staged into alloc_req->host by
seqjob/parjob_prepare_allocation(), so the old code always read a zero and
silently never enforced the cap.
Now both the read source and write destination are derived independently
from each resource's own XCE_RESF_GLOBAL flag (vres->res for the source,
vres->vres for the destination) -- the same flag
calc_wanted_global_resvals()/calc_wanted_host_resvals() (resreq.c) already
dispatch on, so this is the existing rule applied consistently rather than
a new one. Also drops sched_fill_limit_requests()'s unused xce_scheduler_t*
parameter.
Full scheduler suite: 76/77 (the one failure, test_job_rusage_reconciles_
mem_reservation_against_usage, is bug-1762, pre-existing and unrelated).
Live-verified against a fresh single-node deploy.
Filed bug-1790 to track a known, deliberately out-of-scope residual gap:
a future global limit on a per-host resource other than "slots" would need
calc_wanted_host_resvals() to also run during the global phase for that
resource, which nothing currently does.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GLfB5uWRdmx2WiZSFC7Fm7
While live-verifying gitea#46, found and fixed a cascade of 5 previously
undetected bugs that together made every usergroup-scoped limit in
limits.yaml silently non-functional:
- Named usergroups/hostgroups (users.yaml/hosts.yaml) were never
registered into the scheduler's group managers (groupmgr_add() had zero
production callers), so any limit referencing one by name could never
resolve real membership.
- An unconfigured limit scope (e.g. no "projects:") arrived as an
empty-but-present capnp group, not NULL, so build_consumer() built a
real empty group (matching no one) instead of falling back to "all".
- Three separate call sites conflated key_find()'s negative "not found"
return with a real error, short-circuiting before ever reaching the
"instantiate on first use" logic right below them -- the same defect
class already fixed once before in this codebase (bug-1495).
- Two call sites passed &(sched->core.resmgr) where a plain handle was
expected, one level of indirection too many, causing RTTI's key_find()
to reject it as an invalid handle.
All five are live-verified: fresh cluster deploys now correctly match and
instantiate usage tracking for all 7 limits in limits.yaml on every job
submission, reproduced across multiple independent from-scratch redeploys.
A 6th, architecturally distinct issue remains open (bug-1789, logged, not
fixed): sched_fill_limit_requests() computes a job's consumption of a
slots-scoped global limit by reading the wrong request vector, so the
numeric cap itself still never binds. This needs a design decision on
intended cross-scope semantics, not a mechanical fix, so it's left for a
follow-up.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Hb6Edtj4KbS8CrgEWYdA2E
Add --jobs N to xce-build.sh: for meson compile it's passed as ninja's -l N
load-average ceiling (re-checked before every job launch for the whole
build, so it adapts if a concurrent session in this shared sandbox loads
the system mid-build), and as a job-count ceiling (-j/--num-processes,
-p) for meson test/go test, which have no load-average equivalent. No
default -- omitting it preserves prior auto-parallelism behavior exactly.
The xce-build skill now computes N live at invocation time from nproc,
/proc/meminfo's MemAvailable, and /proc/loadavg, rather than a fixed
constant, and passes it through.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01DrxU1dSRYNMV9CDTXHFtcR
ControlJob() set xce_jobctrl_t.action to the job_ctrl_type_t action code
(kill/stop/resume/checkpoint), but that field is actually the capnp
JobControl_u_which union discriminant -- a different enum entirely. This
made kill/resume/checkpoint hit request.c's -EINVAL default case, and made
stop collide with the checkpoint discriminant and read through the wrong
union member. Live-verified against a real deployed cluster: kill/stop/
resume/checkpoint now behave correctly with no apiserver crash.
Also additively accept a human-readable job ID ("{cluster}-{apiserver}-
{timestamp}-{seq}") on the job control endpoint, alongside the existing
numeric form xcecli already translates to -- lets a caller hit the REST
API directly without reimplementing xcecli's ID packing. Live-verified via
a direct curl call with the human-readable ID.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Hb6Edtj4KbS8CrgEWYdA2E
modelJobToSubmit() never set job_startinfo_t.execuser when converting
a submitted job to the C struct sent to the controller/worker, so
every job ran as the worker's 'nobody' default and failed to exec
(exit 126). Default execuser to the submitting user via a new
resolveExecUser() pure helper, and free the newly allocated C string.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Hb6Edtj4KbS8CrgEWYdA2E
sim_reconcile_unaccounted_jobs() (the mechanism that replays historical
DISPATCH decisions in --sim mode) reconstructed a job's allocation but
never called the picker's on_dispatch() hook, unlike the live allocator
path which always does. The job therefore stayed a live candidate in the
picker's own pendJobs bookkeeping, and sim_picker_thread_fn() (which
unconditionally PENDs whatever it is still offered) kept re-offering and
re-PENDing it every subsequent round -- live-observed as an already-
dispatched job receiving a bogus PEND decision (bug-1759: 22/50 jobs in
one batch; bug-1779: 5/10 array-job elements).
This directly contradicts scheduler-sim-decision-parity's own design.md,
which already documented "the picker never re-decides an already-resolved
job" as the correct assumption -- the implementation had drifted from it.
Fixed by calling on_dispatch() immediately after reconstructing the job's
allocation, mirroring sched_allocator()'s own live-dispatch ordering.
test_sim_verify_no_false_mismatch_across_round_separated_dispatch's own
fixture had baked in the stale-PEND symptom as "expected" data; updated to
match the corrected behavior (dispatch-only round 1, empty filler rounds).
Both design docs updated in place with correction notes.
info worker's assigned-jobs listing gated only on job->allocs != NULL,
so a job already at a terminal state (DONE/EXIT) could still appear as
"assigned" if its allocs pointer was ever left stale -- live-observed via
a real captured replay, three DONE jobs kept showing as assigned to a
worker despite the worker's own slot-usage accounting correctly excluding
them. Root cause of the stale allocs itself is not yet pinned (suspected
snapshot save/load inconsistency); this adds the same defensive "terminal
is permanent" exclusion job_state_regresses_terminal()/sim_reconcile_
unaccounted_jobs() already apply elsewhere, at the display layer.
NJOBS/RUN/SSUSP/USUSP (xcecli hosts list -l) and EXEC_HOST (xcecli jobs
list) always showed a hardcoded 1 slot per host regardless of a job's
actual per-host allocation -- rawJobToModel()'s ExecHosts JSONB builder
never read the real per-host "slots" resource from the C job allocation,
so the already-correct downstream slot-counting logic (computeHostSlotUsage,
bug-1743) and EXEC_HOST's "N*host" formatter never had real data to work
with. convert.go now extracts each host's real slots allocation via a new
CGO indexing helper, mirroring the scheduler's own get_nslots_in_alloc()
convention.
Also wires hosts list -l's "CURRENT LOAD USED FOR SCHEDULING" table (tmp/
swap/mem/slots and friends) to the host's real live resource values --
previously entirely hardcoded placeholder strings, never connected to any
backend data.
bug-1554: init_host_attrs() replaces hardcoded placeholder stubs with real
detection -- maxswap from sysinfo()'s totalswap (reusing the same syscall
snapshot already taken for maxmem), maxtmp from a fresh statvfs("/tmp")
total-capacity measurement. No synthetic fallback on detection failure, so
a system with swap genuinely disabled correctly reports 0.
bug-1776: the scheduler simulator's "info worker" assigned-jobs listing now
annotates a job consuming more than one "slots" unit (e.g. "jobid*2"),
using per-host allocation data that was already being walked but discarded.
Falls back to the pre-existing per-job listing when jobs on the same worker
have differing slot counts, since a single suffix on a compressed range
would be ambiguous.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Hb6Edtj4KbS8CrgEWYdA2E
test_info_job_reports_state_queue_priority now expects "Tasks: min =
1, max = 1" instead of the pre-fix "min = 0, max = 0" incidental
byproduct (a fixture's parse_select() failing on an unregistered
"type" resource used to leave the primary's task-count fields at
memset-zero; apply_job_task_count() now correctly overwrites them to
the job's real submitted default of 1/1). This was the one deferred
item tracked in tasks.md's Implementation Notes since task 2.1.
Full-suite regression confirmed clean: scheduler/controller/worker
suites pass except the pre-existing, unrelated bug-1758 (predates
this feature, attributed to a separate spec). A full /xce-build all
also surfaces bug-1723, a pre-existing lib-suite fixture-teardown
leak dating to before this feature and outside Requirement 5.1's
named scope (scheduler/controller/worker) -- confirmed untouched by
any of this feature's commits.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01XYdtrEWWZfKRRGED7raEYg
test_parjob_e2e_multislot.c's standing "bug-1770" assertion now
expects the correct, real primary->minTasks/maxTasks (1/JOB_MAXTASKS)
instead of the buggy always-1 value it used to document. The
PATCH_MAXTASKS_FOR_BUG_1770 block that manually overwrote the
already-ingested resreq's task-count fields is removed entirely --
the full multi-host allocation scenario (2-host commit, per-host
slots) now runs end to end through unmodified real ingest. Header
comment updated to record that bug-1770 is fixed by this feature
rather than describing a live gap.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01XYdtrEWWZfKRRGED7raEYg
Extends test_maxtasks_wiring_ingest.c with two tests: (1) a resource-
request update never changes a job's already-dispatched task count,
because core_events.c's live JOB_UPDATE handler never rebuilds
job->schedobj (Requirement 1.4); (2) two otherwise-identical jobs
differing only in requested task count don't collide in the
unsatisfiable-resource-request cache on the fresh-ingest path
(Requirement 5.4).
Adds a new file, test_maxtasks_wiring_sim_driver_parity.c, proving the
simulation driver's real static translator chain
(sim_process_replay_event_message()/replay_event_to_core_event())
applies this feature's wiring identically to the real driver -- a
genuine white-box test of the real sim_loop.c code (following
test_sim_event_loop.c's established #include technique), not a mirror
of its logic. This was split into its own file because
test_maxtasks_wiring_ingest.c already links the full scheduler_srcs
object list for its own task-3.1 test, which collides at link time
with white-box #include-ing core_events.c/sim_loop.c in one
translation unit; see tasks.md's Implementation Notes.
No production code changed.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01XYdtrEWWZfKRRGED7raEYg
Adds test_maxtasks_wiring_ingest.c proving: (1) fresh ingest's
dispatched primary resreq reflects a real multi-task job's requested
count via the real core_build_job()/JOBS_PUSH path; (2) fresh ingest
and snapshot-restore ingest derive the identical task count for an
equivalent job, asserted both against expected constants and via
direct cross-comparison (Requirement 5.6); (3) a single-task job
reaches identical real host-selection/allocation behavior through the
live allocator dispatch chain, proving the feature is a true no-op
for maxTasks==1 (Requirements 3.1/3.2). No production code changed --
tasks 2.1/2.2 already wired the fix; this adds cross-call-site and
behavioral proof neither of those isolated tests attempted.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01XYdtrEWWZfKRRGED7raEYg
Adds white-box tests for job_start_remote_task()'s launch-type gate
(MPI types defer to their own remote-launch mechanism; default type
fans out via xcecli_runtask() across every allocated host, verified
via a #define-before-#include mock capturing taskrun/taskhost
contents) and create_mpi_cmdline()'s host-list/command construction
for openmpi and mpich. Zero production code changes -- this closes a
previously zero-coverage gap in existing, unmodified worker code
(Requirement 4.4), independent of this feature's resreq wiring work.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01XYdtrEWWZfKRRGED7raEYg
snapshot_load.c's job-restore path now calls apply_job_task_count()
with the job's already-restored mintasks/maxtasks, right after its
resreq parse succeeds and before snapshot_build_schedobj() reads the
resreq's task count for its fingerprint. This is the simulation/replay
driver's only live call site (xce_sched_core_load_snapshot(), reached
via --sim), independent of and parallel to task 2.1's fresh-ingest
wiring in core_events.c.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01XYdtrEWWZfKRRGED7raEYg
core_build_job() now calls apply_job_task_count() right after its
resreq parse succeeds and before core_build_schedobj() reads the
resreq's task count for resreq_fp — so a freshly-submitted multi-task
job's dispatched primary resreq reflects its real requested min/max
task count instead of the previous always-one-task hardcode.
Tracked follow-up (not blocking, see tasks.md Implementation Notes):
test_repl_breakpoints_inspection.c:735 hardcodes an incidental
"Tasks: min = 0, max = 0" byproduct of a fixture's parse_select()
failure; now correctly reports (1,1) per Requirement 1.2. Fix folded
into task 5.1's full-suite regression pass.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01XYdtrEWWZfKRRGED7raEYg
Adds apply_job_task_count() to overwrite a standalone primary resreq's
minTasks/maxTasks with a job's real requested task count, no-op for
COMPOUND/ALTER nested primaries, with a defensive min>max clamp.
Closes the gap that lets scheduler-allocation's multi-host allocator
ever be reached (bug-1770). Wiring at live ingest call sites follows
in subsequent tasks.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01XYdtrEWWZfKRRGED7raEYg
Drive a real multi-slot parallel job through the actual production
dispatch path -- real sched_init(), real queue/worker/job ingest via
xce_core_push_events(), a real scheduling session (exercising task 5.2's
cache-clear wiring), the real sched_allocator() thread, cache lookup,
parjob_allocator's real my_alloc(), host_iter, and commit -- and confirm
the resulting job allocation genuinely spans 2 hosts with correct
per-host slot counts (Req 4.1, 4.2, 4.3).
Discovered and logged bug-1770: parse_resreq_primary() unconditionally
hardcodes primary->minTasks/maxTasks to 1 regardless of what a job
actually requests, and nothing in the real ingest pipeline ever copies
the job's own submitted task count (correctly threaded from the
apiserver's num_processors/maxnum_processors all the way to
xce_sched_job_t.mintasks/maxtasks) onto the nested xce_primary_resreq_t
fields the allocator dispatch actually reads. This predates the feature
by ~7 months and is out of this spec's scope to fix, but it means
parjob_allocator has zero live production path until a dedicated
follow-up lands. The test compensates with one clearly-labeled patch of
the real, already-ingested resreq's task-count fields immediately after
real parsing; everything else in the pipeline is unmodified production
code.
Also logged bug-1771 (xce_sched_fcfs's fcfs_del_job() dereferences
job->queue->name with no NULL guard) discovered while building this
task's fixture; worked around by registering the job's queue before
push, mirroring the apiserver's own pre-submission queue validation.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01XYdtrEWWZfKRRGED7raEYg
Clear the cache inside core_begin_session_internal() so a job cached as
unsatisfiable in one session is re-evaluated fresh in the next (Req 7.2).
An initial weak-symbol cast-based approach was rejected: it corrupted
xce_sched_sim_driver_t.alt_events_file, since core_begin_session_internal()
is shared between driver.c (real xce_scheduler_t) and sim_loop.c (a
structurally different, deliberately isolated sim driver struct with no
cache field at the same byte offset) -- the strong cache_clear() symbol
won unconditionally for both, invisible to ASAN and untested by any
test_sim_*.c file.
Fixed via a driver-registered callback: xce_sched_core_t gains a
cache_clear_cb hook (NULL by default, per xce_sched_core_init()'s
full-struct memset), set only in main.c's driver-mode branch --
mirroring the existing on_decision/xce_core_set_decision_cb() pattern.
--sim mode returns before that registration is ever reached, so its hook
stays NULL and core_begin_session_internal()'s guarded call is a no-op.
The real xce_scheduler_t-casting adapter lives in alloc_cache.c rather
than driver.c, since ~11 tests white-box-include driver.c without
linking alloc_cache.c.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01XYdtrEWWZfKRRGED7raEYg
Wrap sched_make_primary_alloc_() -- the single choke point all three
primary-resreq shapes (standalone, compound-nested, alter-nested via
compound delegation) already funnel through -- with a cache lookup
before allocation and a cache insert on exhaustion. On a hit, pend the
job immediately via the same make_pendreason()/add_job_preason() path
every exhaustion case already uses, without invoking either allocator
(Req 7.3, 7.4, 7.5).
A nested primary has no real per-primary schedobj, so
sched_primary_cache_obj_() synthesizes one on the stack per dispatch for
cache-key purposes; verified memory-safe since alloc_cache.c's
build_key_() copies the fingerprint/identity by value synchronously
before the stack frame unwinds.
Log bug-1768 (compound/alter-nested primaries never get real slot/rusage
availability checks -- seqjob/parjob's prepare_allocation hooks no-op
whenever ctx->schedobj->tag != XCE_RESREQ_PRIMARY) and bug-1769
(sched_make_alter_allocation() type-confuses xce_compound_schedobj_t*
as xce_sched_obj_t*, a ~112-byte heap over-read, currently untested).
Both are pre-existing and out of this feature's scope; discovered and
confirmed while building this task's dispatch tests.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01XYdtrEWWZfKRRGED7raEYg
Task 4.3's multi-host merge itself was already implemented and verified
by task 4.2. The one unproven clause -- job-level shared-resource
allocation computed once regardless of host count -- is now covered by
test_shared_resource_debited_once_regardless_of_host_count, using a
PER_JOB (not PER_TASK) global resource so a per-bucket-repeated
computation can't coincidentally sum to the correct total. Mutation
testing (a deliberately injected duplicate per-host shared debit)
confirms the test fails exactly as expected; production code required
no change.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01XYdtrEWWZfKRRGED7raEYg
Replace parjob's return-0 stub with a real multi-host allocator: for each
candidate plan, walk buckets in the given (descending) order, claiming a
distinct host per bucket via a fresh host_iter excluding hosts already
claimed elsewhere in the same plan attempt. On a bucket failure, reverse
every earlier successful claim's provisional debit via
sched_ctx_uncommit_host() before advancing to the next plan (Req 4.2, 4.4,
4.5, 4.6, 4.7).
Fix parjob_prepare_allocation() to compute each bucket's want from that
bucket's own slot count (ctx->nslots) instead of a fixed value, and reset
ctx->alloc_req.host at the top of each bucket's scan so host_iter's
pre-filter can't inherit a stale want from a differently-sized prior
bucket (safe: sched_try_alloc_on_host_() always repopulates it from the
real per-bucket want before its own authoritative check runs).
Add a maxTasks<=1 guard matching Req 4 AC1's precondition, needed because
a snapshot-restored job with unpopulated maxTasks==0 (pre-existing,
tracked separately as bug-1767) would otherwise reach parjob's now-real
allocator instead of being silently ignored as before.
On full success, merges every claimed bucket into one xce_job_alloc_t via
make_job_alloc()'s existing runtime-sized host array, satisfying task
4.3's multi-host commit scope as well.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01XYdtrEWWZfKRRGED7raEYg
Add the real xce_alloc_plan_iter_t for parjob: for host-count k=1..maxTasks
(ascending, most-preferred-first), generate one balanced partition of
maxTasks into k buckets (floor/ceil split, remainder to the first rem
buckets), which is descending by construction. Pure function of maxTasks
alone -- never dereferences sched -- with a real pthread_testcancel()
cancellation point and stable post-exhaustion NULL (Req 2.1-2.6, 4.1).
my_alloc()/parjob_prepare_allocation() remain untouched stubs; wiring the
iterator into real allocation is task 4.2.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01XYdtrEWWZfKRRGED7raEYg
Add a dedicated equivalence suite (test_seqjob_rewrap_equivalence.c)
covering multi-host fixtures the trivial single-host test doesn't reach:
select-filter-only rejection, resource-only rejection, a combined
select-then-resource-then-success ordering, and per-host debit
correctness. The iteration-order test is grounded in the deleted
pre-rewrap sched_collect_workers() code path (verified via git show
against the parent commit) and the unmodified xlist.c primitive, not
merely in host_iter.c's own traversal, so it proves equivalence to the
old behavior rather than internal self-consistency.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01XYdtrEWWZfKRRGED7raEYg
Restructure seqjob's my_alloc() as a trivial single-bucket case over the
shared xce_alloc_plan_iter_t/xce_host_iter_t contracts, replacing its own
inline worker-collection loop. Guard seqjob_prepare_allocation() to
maxTasks == 1, closing the fan-out gap for the parallel allocator's later
fix (Req 9.1, 4.6).
Export sched_try_alloc_on_host_() from resource.c (drop static/va_list,
plain (worker, sched, ctx) signature) so seqjob calls the real per-host
try-alloc leaf instead of a private duplicate; delete the now-dead
sched_try_alloc_on_hosts() variadic wrapper. Also wires host_iter.c into
the production scheduler_srcs list (task 2.1 only added it to testsuite).
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01XYdtrEWWZfKRRGED7raEYg
Adds alloc_cache.c: xce_cache_key_t, cache_key_ops (the corrected
three-way, 0-means-equal .cmp -- an inverted-boolean version was a
real bug found during design review), and cache_lookup/insert/clear
against a real kvtab_t (constructed matching xce_select.c's
sel_node_table_init() precedent). Cache values are a boxed reason-
category int, not a bare hit/miss sentinel, satisfying Req 7 AC5's
requirement that a hit reproduce the original pend-reason category.
unsatisfiable_cache lives on xce_scheduler_t per Req 7 AC1.
Task 2.2 (scheduler-allocation spec). Requirements 7.1, 7.3, 7.4, 7.5.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01XYdtrEWWZfKRRGED7raEYg
Adds host_iter.c: the one shared xce_host_iter_t implementation,
lazily filtering sched->core.workers by select expression
(sched_meet_host_selection, unchanged) and resource availability
(sched_check_host_satisfy, unchanged), with a real cancellation
point per candidate visited. xce_alloc_plan_t gains a ctx field
(xce_sched_ctx_t*) so the iterator can reach the scheduler/select
expression/staged resource-want -- additive to task 1.1's approved
interface, NULL-guarded at init(). Test proves the core design
property this spec exists to establish: a host too small for one
search remains a valid candidate for a later, smaller search via a
fresh iterator instance over the same worker set.
Task 2.1 (scheduler-allocation spec). Requirements 3.1-3.6.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01XYdtrEWWZfKRRGED7raEYg
Adds sched_ctx_uncommit_host() (resource.c), the structural inverse
of sched_try_alloc_on_host_()'s existing XVEC_ADD_FLOAT commit call,
scoped to ctx->uncomitted only per requirements.md's Rollback scope
discussion. Test exercises the real sched_ctx_get_uncomit_host()
(schedobj.c) via white-box #include rather than a private
reimplementation, matching test_picker_config_fields.c's established
precedent.
Task 1.4 (scheduler-allocation spec). Requirement 4.2.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01XYdtrEWWZfKRRGED7raEYg
Adds xce_alloc_mode_t (XCE_ALLOC_NORMAL=0, XCE_ALLOC_PREEMPT, with
commented-out RESERVE/BACKFILL placeholders) and alloc_mode on
xce_sched_ctx_t, zero-defaulted by sched_ctx_init()'s existing
memset. Purely additive: nothing branches on the field yet -- this
spec only declares the extension point scheduler-preemption needs.
Task 1.3 (scheduler-allocation spec). Requirement 9.5.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01XYdtrEWWZfKRRGED7raEYg
Adds rusage_hash to xce_primary_resreq_t (computed once by
parse_resreq_primary() via a new hash_rusages() helper, hashing
rusage content rather than raw struct bytes since resname is a
non-interned strdup'd pointer), the xce_resreq_fingerprint_t value
type, and resreq_fp on xce_primary_schedobj_t, populated once at real
job-ingest time in core_build_primary_schedobj().
Task 1.2 (scheduler-allocation spec). Requirement 7.1.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01XYdtrEWWZfKRRGED7raEYg
Defines xce_alloc_plan_t (partition of a job's task count into
per-host slot buckets), xce_alloc_plan_iter_t, and xce_host_iter_t
(init/next/fini vtables, stack-allocate-then-init-in-place convention
matching sched_ctx_init) as reusable abstractions any allocator can
adopt. Proven usable by a trivial no-op reference implementation and
a fresh-per-bucket host_iter composition test, no allocator wired in
yet.
Task 1.1 (scheduler-allocation spec). Requirements 2.1, 3.1.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01XYdtrEWWZfKRRGED7raEYg
An idle scheduler (nothing queued to publish) never noticed a dead
controller connection: sched_needs_rejoin was only ever set as a side
effect of an active client-library call, and an idle round makes none.
After any controller failover with nothing already in flight, jobs
submitted to the new primary stayed PEND indefinitely.
Adds xcecli_is_connection_broken() (libxce) -- a pure, socket-free read
of the transport's already-async-maintained dead/alive flag, set for
free by the existing RDHUP disconnect handler with zero application
calls. sched_check_pending_rejoin() now retries xcecli_join() on either
the existing reactive flag or this new passive signal, so a healthy
idle scheduler still makes zero extra network calls while a genuinely
dead connection self-heals within one ~1s round -- live-verified
against a 3-node HA cluster (rejoin fired ~300ms after new-primary
election with nothing submitted in between).
Also closes out bug-1735 (job memory/CPU usage reporting): confirmed
already fixed on this branch since 2026-08-30, just never marked in
the bug log.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01XYdtrEWWZfKRRGED7raEYg
The terminal step: src/frontend/Makefile (194 lines, the second,
independently-maintained build/test path this whole spec exists to
eliminate) is deleted. Re-ran the full task 4.1 validation suite with
the file genuinely absent (build, meson test --suite apiserver, and
-- going beyond citing task 3.5 -- a real live run-resource-verify.sh
end-to-end run: VERDICT result=PASS, closing design.md's Testing
Strategy "Discovery-path regression" check specifically). Full-repo
sweep: 181 hits, zero unaccounted, byte-identical to task 4.1's
non-Makefile hits (the only delta is the Makefile's own 8
self-referential lines, now gone since the file doesn't exist).
Independently re-verified by the task reviewer.
Meson is now the single build/test system for apiserver and xcecli.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01XYdtrEWWZfKRRGED7raEYg
Comment-only edits in 4 files: test_ha_bootstrap.c/test_ha_ctrlhello.c
rename their "make test-integration" pointers to reference the
regression suite (tests/regression/run.sh) directly; apiserver/xcecli
cmd/main.go drop the Makefile clause from their ldflags-injection
comments, keeping only src/tools/go_build. Zero production/test
logic changed -- verified via full controller unit-test suite rerun
(1261 tests, 0 failures) and xcecli build.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01XYdtrEWWZfKRRGED7raEYg
