lei.chen/myxce
1
0
Fork 1
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

fix(scheduler): gate resource value/used display on XCE_RESF_CONSUMABLE

Browse source 
repl_info_worker_one() ("info worker <name>") unconditionally showed a
total=/available= pair for every numeric per-worker resource, regardless
of whether it was actually consumable -- misleading for a static
per-worker attribute nothing ever allocates against.

repl_info_resource_one() ("info resource <name|all>") never reported a
consumable GLOBAL/SHARED resource's in-use amount at all, even though
core->sharedInUse (the cluster-wide mirror of a worker's resInUse)
already exists and is populated by the real resource-accounting code.

Both now gate on XCE_RESF_CONSUMABLE: consumable resources report
value=<capacity> used=<in-use>, non-consumable resources keep a single
value=<current> line. Also added the resource definition's missing
ordtype= field.

Add test_info_resource_reports_consumable_value_and_used and
test_info_worker_reports_consumable_value_and_used; update 3 pre-existing
tests whose asserted output shape changed.

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
This commit is contained in:
Rongsong Shen 2026-08-20 07:27:00 +00:00
parent 51063c1dc8
commit 9473361683
No known key found for this signature in database
GPG key ID: 5BCDAAE0123E9FE2
2 changed files with 250 additions and 48 deletions
Download patch file Download diff file Expand all files Collapse all files

143
src/scheduler/src/repl_info.c
View file

@ -278,18 +278,21 @@ static void repl_info_queue(xce_sched_sim_driver_t* sim, const char* arg)
/*
* repl_info_worker_one - AC 22's "info worker <name>": "Dump one worker:
* state, resources (total + available), assigned jobs." `w->resvals` holds
* currently-AVAILABLE values (the same zarray the picker/allocator
* accounting path reads/resets each session, core_events.c); `w->resInUse`
* holds the currently-committed amount at the same index -- "total" below
* is therefore computed as available+inUse, mirroring how this codebase's
* own resource accounting treats the pair as a conservation invariant
* (nothing elsewhere sums them into a literal "total" field to reuse).
* Resource names are resolved via find_resource_name_by_idx() against a
* resmgr walk built the same way xce_sched_core_save_snapshot() builds its
* own `res_arr` (snapshot.c) -- not a new lookup mechanism. "assigned jobs"
* scans core->jobs for any per-host allocation naming this worker
* (xce_job_alloc_t.allocs[]->host, xce/types.h).
* state, resources, assigned jobs." `w->resvals` holds currently-AVAILABLE
* values (the same zarray the picker/allocator accounting path reads/resets
* each session, core_events.c); `w->resInUse` holds the currently-committed
* amount at the same index. Only a resource flagged XCE_RESF_CONSUMABLE has
* a capacity/in-use split worth reporting -- its value= is available+inUse
* (a conservation invariant this codebase's own resource accounting already
* relies on) and used= is inUse; a non-consumable numeric resource (a
* static per-worker attribute nothing ever allocates against) gets a single
* value= line instead, mirroring repl_info_resource_one()'s identical
* CONSUMABLE gate for the cluster-wide case. Resource names/flags are
* resolved against a resmgr walk built the same way
* xce_sched_core_save_snapshot() builds its own `res_arr` (snapshot.c) --
* not a new lookup mechanism. "assigned jobs" scans core->jobs for any
* per-host allocation naming this worker (xce_job_alloc_t.allocs[]->host,
* xce/types.h).
*/
static void repl_info_worker_one(xce_sched_sim_driver_t* sim,
xce_sched_worker_t* w)
@ -319,27 +322,34 @@ static void repl_info_worker_one(xce_sched_sim_driver_t* sim,
}
for (i = 0; i < w->resvals.count; i++) {
xce_val_t* avail_v = (xce_val_t*)zarray_get(&(w->resvals), i);
xce_val_t* used_v = (xce_val_t*)zarray_get(&(w->resInUse), i);
const char* rname = find_resource_name_by_idx(res_arr, res_n, i);
xce_val_t* avail_v = (xce_val_t*)zarray_get(&(w->resvals), i);
xce_val_t* used_v = (xce_val_t*)zarray_get(&(w->resInUse), i);
const char* rname = find_resource_name_by_idx(res_arr, res_n, i);
xce_resource_t* rdef = NULL;
int j;
if ((avail_v == NULL) || (rname == NULL)) {
continue;
}
for (j = 0; j < res_n; j++) {
if (res_arr[j]->idx == i) {
rdef = res_arr[j];
break;
}
}
/* bug-1590: xce_val_t.which (a Value_u_which wire tag,
* controller.capnp.h) tells us the value's REAL type -- bug-1568
* fixed comparing it against the wrong enum for the numeric case,
* but left every non-numeric resource (Value_u_svalue/
* Value_u_bvalue -- e.g. this worker's `type`/`model` string
* resources, or any boolean-typed resource) silently coerced to
* 0.0 by the old unconditional %.2f formatting below. "total"/
* "available" (a consumable-quantity concept: how much of this
* worker's capacity is free vs. in use) is only meaningful for a
* NUMERIC resource -- a string or boolean resource is a single
* static-ish per-worker attribute with no "in use" counterpart, so
* it gets its own "value=" line instead of a fabricated
* total/available pair. */
* 0.0 by the old unconditional %.2f formatting below. A value/used
* split is only meaningful for a CONSUMABLE numeric resource -- a
* string/boolean resource, or a non-consumable numeric one (a
* static per-worker attribute nothing ever allocates against), gets
* its own single "value=" line instead. */
switch (avail_v->which) {
case Value_u_svalue:
fprintf(stdout, "[sim] resource %s: value=%s\n", rname,
@ -351,17 +361,22 @@ static void repl_info_worker_one(xce_sched_sim_driver_t* sim,
avail_v->u.bvalue ? "true" : "false");
break;
case Value_u_fvalue: {
double avail = avail_v->u.fvalue;
double used = ((used_v != NULL) && (used_v->which == Value_u_fvalue))
? used_v->u.fvalue
: 0.0;
case Value_u_fvalue:
if ((rdef != NULL) && (rdef->flags & XCE_RESF_CONSUMABLE)) {
double avail = avail_v->u.fvalue;
double used = ((used_v != NULL) &&
(used_v->which == Value_u_fvalue))
? used_v->u.fvalue
: 0.0;
fprintf(stdout,
"[sim] resource %s: total=%.2f available=%.2f\n", rname,
avail + used, avail);
fprintf(stdout,
"[sim] resource %s: value=%.2f used=%.2f\n", rname,
avail + used, used);
} else {
fprintf(stdout, "[sim] resource %s: value=%.2f\n", rname,
avail_v->u.fvalue);
}
break;
}
default:
/* ivalue32/uvalue32/ivalue64/uvalue64 -- not a shape this
@ -487,21 +502,27 @@ static void repl_resource_flags_str(int flags, char* buf, size_t bufsz)
/*
* repl_info_resource_one - AC 22 amendment (task 12.5, bug-1593): "info
* resource <name>" reports the resource DEFINITION's own metadata (desc,
* value type, XCE_RESF_* flags, reporting interval) -- never a per-host
* aggregation, which is repl_info_worker_one()'s job. A resource flagged
* XCE_RESF_GLOBAL or XCE_RESF_SHARED additionally gets its current value,
* read directly from the core's single defining pool (`core->sharedInst`,
* indexed by `r->idx` -- the same pool RESVALS_FULL/RESVALS_DELTA event
* handling and snapshot save/load already treat as authoritative for these
* resources, core_events.c/snapshot.c). A host-based resource (the common
* case -- cpus, mem, ...) has no single cluster-wide value to report here;
* its per-host values remain `info worker <name>`'s job.
* value type, order type, XCE_RESF_* flags, reporting interval) -- never a
* per-host aggregation, which is repl_info_worker_one()'s job. A resource
* flagged XCE_RESF_GLOBAL or XCE_RESF_SHARED additionally gets its current
* value, read directly from the core's single defining pool
* (`core->sharedInst`, indexed by `r->idx` -- the same pool RESVALS_FULL/
* RESVALS_DELTA event handling and snapshot save/load already treat as
* authoritative for these resources, core_events.c/snapshot.c). A host-based
* resource (the common case -- cpus, mem, ...) has no single cluster-wide
* value to report here; its per-host values remain `info worker <name>`'s
* job. If the resource is additionally XCE_RESF_CONSUMABLE, the numeric
* value line also reports how much is currently in use, read from
* `core->sharedInUse` at the same index (mirrors repl_info_worker_one()'s
* resInUse-derived split below) -- a non-consumable numeric resource has no
* "used" counterpart, so it keeps a single value= line.
*/
static void repl_info_resource_one(xce_sched_sim_driver_t* sim,
xce_resource_t* r)
{
char flags_str[160];
const char* valtype_str;
const char* ordtype_str;
switch (r->valtype) {
case XCE_VAL_STRING:
@ -515,12 +536,27 @@ static void repl_info_resource_one(xce_sched_sim_driver_t* sim,
valtype_str = "numeric";
break;
}
switch (r->ordtype) {
case XCE_ORD_INCR:
ordtype_str = "incr";
break;
case XCE_ORD_DECR:
ordtype_str = "decr";
break;
case XCE_ORD_NA:
default:
ordtype_str = "na";
break;
}
repl_resource_flags_str(r->flags, flags_str, sizeof(flags_str));
fprintf(stdout,
"[sim] resource %s: desc=%s valtype=%s flags=%s interval=%d\n",
"[sim] resource %s: desc=%s valtype=%s ordtype=%s flags=%s "
"interval=%d\n",
r->name, (r->desc != NULL) ? r->desc : "(none)", valtype_str,
flags_str, r->interval);
ordtype_str, flags_str, r->interval);
if ((r->flags & (XCE_RESF_GLOBAL | XCE_RESF_SHARED)) == 0) {
fprintf(stdout, "[sim] host-based resource -- see 'info worker "
@ -551,7 +587,28 @@ static void repl_info_resource_one(xce_sched_sim_driver_t* sim,
v->u.bvalue ? "true" : "false");
break;
case Value_u_fvalue:
fprintf(stdout, "[sim] value=%.2f\n", v->u.fvalue);
/* A consumable resource has a capacity/in-use split worth
* reporting; a non-consumable numeric resource (e.g. a static
* cluster-wide attribute) has no "used" counterpart, so it gets
* a single value= line like the string/boolean cases above. */
if (r->flags & XCE_RESF_CONSUMABLE) {
xce_val_t* used_v = NULL;
double used = 0.0;
if ((r->idx >= 0) &&
(r->idx < sim->core.sharedInUse.count)) {
used_v = (xce_val_t*)zarray_get(&(sim->core.sharedInUse),
r->idx);
}
if ((used_v != NULL) && (used_v->which == Value_u_fvalue)) {
used = used_v->u.fvalue;
}
fprintf(stdout, "[sim] value=%.2f used=%.2f\n",
v->u.fvalue + used, used);
} else {
fprintf(stdout, "[sim] value=%.2f\n", v->u.fvalue);
}
break;
default:
fprintf(stdout, "[sim] value=(unrecognized wire type)\n");

155
src/scheduler/testsuite/test_repl_breakpoints_inspection.c
View file

@ -732,11 +732,12 @@ static void test_info_worker_reports_resource_values(void)
captured = capture_dispatch(&sim, "info worker hostA");
TEST_ASSERT_NOT_NULL_MESSAGE(
strstr(captured, "resource mem: total=42.50 available=42.50"),
strstr(captured, "resource mem: value=42.50"),
"a worker's real numeric resource value must be displayed, not "
"the XCE_VAL_NUMERIC/Value_u_fvalue type-tag mismatch's 0.00 "
"(bug-1568)");
TEST_ASSERT_NULL_MESSAGE(strstr(captured, "total=0.00 available=0.00"),
"(bug-1568); a non-consumable resource (no flags set here) gets a "
"single value= line, not a value=/used= pair");
TEST_ASSERT_NULL_MESSAGE(strstr(captured, "value=0.00"),
"must not fall back to the zero default for a "
"genuinely-set numeric value");
@ -745,6 +746,82 @@ static void test_info_worker_reports_resource_values(void)
printf("test_info_worker_reports_resource_values: PASS\n");
}
/*
* New requirement (2026-08-20): a per-worker resource additionally flagged
* XCE_RESF_CONSUMABLE must report both value (available+inUse) and used
* (inUse); a non-consumable numeric resource keeps a single value= line
* (test_info_worker_reports_resource_values above covers that case).
*/
static void test_info_worker_reports_consumable_value_and_used(void)
{
xce_sched_sim_driver_t sim;
xce_core_event_t events[1];
xce_resource_t wire_resource;
xce_resource_t* r;
xce_worker_t wire_worker;
xce_worker_t* worker_ptrs[1];
xce_val_t val;
xce_name_val_t nv;
xce_name_val_t* nv_ptrs[1];
xce_val_t used_val;
xce_sched_worker_t* sw;
char* captured;
int rc;
memset(&sim, 0, sizeof(sim));
rc = xce_sched_core_init(&(sim.core));
TEST_ASSERT_EQUAL_INT(0, rc);
memset(&wire_resource, 0, sizeof(wire_resource));
wire_resource.name = "slots";
wire_resource.valtype = XCE_VAL_NUMERIC;
wire_resource.flags = XCE_RESF_CONSUMABLE;
r = makeResource(sim.core.resmgr, &wire_resource);
TEST_ASSERT_NOT_NULL(r);
memset(&val, 0, sizeof(val));
val.which = Value_u_fvalue;
val.u.fvalue = 5.0;
memset(&nv, 0, sizeof(nv));
nv.name = "slots";
nv.val = &val;
nv_ptrs[0] = &nv;
memset(&wire_worker, 0, sizeof(wire_worker));
wire_worker.name = "hostA";
wire_worker.status = 0;
wire_worker.nresvals_to_sched = 1;
wire_worker.resvals_to_sched = nv_ptrs;
worker_ptrs[0] = &wire_worker;
memset(events, 0, sizeof(events));
events[0].type = XCE_CORE_EVENT_WORKERS_PUSH;
events[0].u.workers_push.nworkers = 1;
events[0].u.workers_push.workers = worker_ptrs;
rc = xce_core_push_events(&(sim.core), events, 1);
TEST_ASSERT_EQUAL_INT(0, rc);
sw = sched_get_worker(&(sim.core), "hostA");
TEST_ASSERT_NOT_NULL(sw);
memset(&used_val, 0, sizeof(used_val));
used_val.which = Value_u_fvalue;
used_val.u.fvalue = 2.0;
zarray_put(&(sw->resInUse), r->idx, dup_val(&used_val));
captured = capture_dispatch(&sim, "info worker hostA");
TEST_ASSERT_NOT_NULL_MESSAGE(
strstr(captured, "resource slots: value=7.00 used=2.00"),
"a CONSUMABLE worker resource must report value=avail+used and "
"used=inUse");
free(captured);
printf("test_info_worker_reports_consumable_value_and_used: PASS\n");
}
/*
* bug-1590: repl_info_worker_one() forced EVERY resource value through the
* numeric "total=%.2f available=%.2f" format, silently coercing any
@ -893,7 +970,8 @@ static void test_info_resource_reports_string_value(void)
TEST_ASSERT_NOT_NULL_MESSAGE(
strstr(captured, "resource license_pool: desc=(none) "
"valtype=string flags=GLOBAL interval=0"),
"valtype=string ordtype=incr flags=GLOBAL "
"interval=0"),
"a GLOBAL resource's definition metadata must be printed "
"(bug-1593)");
TEST_ASSERT_NOT_NULL_MESSAGE(
@ -912,6 +990,71 @@ static void test_info_resource_reports_string_value(void)
printf("test_info_resource_reports_string_value: PASS\n");
}
/*
* New requirement (2026-08-20): a GLOBAL/SHARED resource additionally
* flagged XCE_RESF_CONSUMABLE must report both its value (capacity,
* available+inUse) and how much is currently in use, read from
* core->sharedInUse -- the cluster-wide mirror of repl_info_worker_one()'s
* resInUse-derived value/used split.
*/
static void test_info_resource_reports_consumable_value_and_used(void)
{
xce_sched_sim_driver_t sim;
xce_core_event_t events[1];
xce_resource_t wire_resource;
xce_resource_t* r;
xce_val_t avail_val;
xce_val_t used_val;
xce_resval_t rv;
xce_resval_t* rv_ptrs[1];
char* captured;
int rc;
memset(&sim, 0, sizeof(sim));
rc = xce_sched_core_init(&(sim.core));
TEST_ASSERT_EQUAL_INT(0, rc);
memset(&wire_resource, 0, sizeof(wire_resource));
wire_resource.name = "licenses";
wire_resource.valtype = XCE_VAL_NUMERIC;
wire_resource.flags = XCE_RESF_SHARED | XCE_RESF_CONSUMABLE;
r = makeResource(sim.core.resmgr, &wire_resource);
TEST_ASSERT_NOT_NULL(r);
memset(&avail_val, 0, sizeof(avail_val));
avail_val.which = Value_u_fvalue;
avail_val.u.fvalue = 8.0;
memset(&rv, 0, sizeof(rv));
rv.name = "licenses";
rv.val = &avail_val;
rv_ptrs[0] = &rv;
memset(events, 0, sizeof(events));
events[0].type = XCE_CORE_EVENT_RESVALS_FULL;
events[0].u.resvals_full.n = 1;
events[0].u.resvals_full.resvals = rv_ptrs;
rc = xce_core_push_events(&(sim.core), events, 1);
TEST_ASSERT_EQUAL_INT(0, rc);
memset(&used_val, 0, sizeof(used_val));
used_val.which = Value_u_fvalue;
used_val.u.fvalue = 3.0;
zarray_put(&(sim.core.sharedInUse), r->idx, dup_val(&used_val));
captured = capture_dispatch(&sim, "info resource licenses");
TEST_ASSERT_NOT_NULL_MESSAGE(
strstr(captured, "value=11.00 used=3.00"),
"a CONSUMABLE global resource must report value=avail+used and "
"used=inUse, sourced from core->sharedInUse");
free(captured);
printf("test_info_resource_reports_consumable_value_and_used: PASS\n");
}
/*
* bug-1593 (task 12.5): `info resource all` must list only resources
* flagged XCE_RESF_GLOBAL/XCE_RESF_SHARED -- a host-based resource (no
@ -1005,7 +1148,7 @@ static void test_info_resource_named_lookup_reports_host_based_note(void)
TEST_ASSERT_NOT_NULL_MESSAGE(
strstr(captured, "resource cpus: desc=(none) valtype=numeric "
"flags=(none) interval=0"),
"ordtype=incr flags=(none) interval=0"),
"a host-based resource's metadata must still be reported by a "
"named lookup (bug-1593)");
TEST_ASSERT_NOT_NULL_MESSAGE(
@ -1661,8 +1804,10 @@ int main(void)
RUN_TEST(test_info_job_unknown_reports_not_found);
RUN_TEST(test_info_queue_reports_status_and_queued_jobs);
RUN_TEST(test_info_worker_reports_resource_values);
RUN_TEST(test_info_worker_reports_consumable_value_and_used);
RUN_TEST(test_info_worker_reports_string_and_boolean_resource_values);
RUN_TEST(test_info_resource_reports_string_value);
RUN_TEST(test_info_resource_reports_consumable_value_and_used);
RUN_TEST(test_info_resource_all_lists_only_global_shared_resources);
RUN_TEST(test_info_resource_named_lookup_reports_host_based_note);
RUN_TEST(test_info_resource_all_reports_when_none_are_global_shared);