feat(worker-rejoin-resync): add reconciliation wire codec helpers (task 1.4)
Add xce_job_state_query_req()/xce_job_state_report_req() to request.c, paginating job IDs / job-state entries across OPCode_jobStateQuery / OPCode_jobStateReport messages via direct Request.body union arms (mirroring xce_jobdispatch_req/xce_notify_req's direct-send shape, not xce_push_data_req's XceData-table multiplexing -- these are dedicated union arms, not multiplexed push/reply-get traffic). 8 round-trip tests (normal/empty/multi-page for both query and report, covering all 4 JobStateStatus values and both queried states) all pass. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01Y4M5MJM2zK6nU43da267Ku
This commit is contained in:
parent
9fbbd12b36
commit
9f3dfe7903
6 changed files with 581 additions and 1 deletions
409
src/lib/testsuite/test_job_state_wire.c
Normal file
409
src/lib/testsuite/test_job_state_wire.c
Normal file
|
|
@ -0,0 +1,409 @@
|
|||
/*
|
||||
* test_job_state_wire.c -- worker-rejoin-resync task 1.4: wire encode/decode
|
||||
* round-trip tests for the job-state reconciliation exchange
|
||||
* (OPCode_jobStateQuery / OPCode_jobStateReport), Req 8.1/8.2/8.3.
|
||||
*
|
||||
* Mirrors test_request.c's HA-heartbeat round-trip style: there is no live
|
||||
* socket in these unit tests, so a real (heap) endpoint_t is built via
|
||||
* endpoint_get() -- NOT a hand-rolled stub struct (bug-378: an
|
||||
* under-sized/wrong-typed stand-in for endpoint_t caused a global-buffer-
|
||||
* overflow when passed where a real endpoint_t* was expected). endpoint_enq()
|
||||
* only queues onto ep->outq (a plain in-memory linked list, dataq_enq()) and
|
||||
* skips the iomgr poll-arm step entirely when ep->iomgr == NULL (the default
|
||||
* after endpoint_get()), so xce_job_state_query_req()/xce_job_state_report_req()
|
||||
* can be exercised end-to-end -- real xce_encode() on send, real xce_decode()
|
||||
* on "receive" -- without a socket, exactly as
|
||||
* test_xce_request_propagates_send_failure() (test_request_send_propagation.c)
|
||||
* already does for a single OPCode_nop request. Each queued page is popped
|
||||
* with endpoint_deq()/ITPTR_TO_PACKET() and decoded with xce_decode(), which
|
||||
* takes ownership of (frees) the packet -- matching test_request.c's
|
||||
* documented xce_decode() contract.
|
||||
*/
|
||||
|
||||
#include <errno.h>
|
||||
#include <stdlib.h>
|
||||
#include <string.h>
|
||||
#include <unity/unity.h>
|
||||
#include "netio.h"
|
||||
#include "request.h"
|
||||
#include "xce/types.h"
|
||||
|
||||
static endpoint_t* new_test_endpoint(void)
|
||||
{
|
||||
endpoint_t* ep = endpoint_get();
|
||||
TEST_ASSERT_NOT_NULL(ep);
|
||||
return ep;
|
||||
}
|
||||
|
||||
/* Pop one queued page off ep->outq and decode it. Fails the test if the
|
||||
* queue is empty. Caller must xce_free_request() the result. */
|
||||
static void pop_and_decode(endpoint_t* ep, xce_request_t* out)
|
||||
{
|
||||
data_item_t* it = endpoint_deq(ep);
|
||||
TEST_ASSERT_NOT_NULL(it);
|
||||
|
||||
io_packet_t* pkt = ITPTR_TO_PACKET(it);
|
||||
|
||||
memset(out, 0, sizeof(*out));
|
||||
TEST_ASSERT_EQUAL_INT(0, xce_decode(pkt, out)); /* consumes/frees pkt */
|
||||
}
|
||||
|
||||
static void assert_queue_empty(endpoint_t* ep)
|
||||
{
|
||||
TEST_ASSERT_NULL(endpoint_deq(ep));
|
||||
}
|
||||
|
||||
/* ── OPCode_jobStateQuery round trips (controller -> worker) ────────────── */
|
||||
|
||||
static void test_job_state_query_req_null_args(void)
|
||||
{
|
||||
xce_job_state_query_t query = { 0 };
|
||||
char nm[] = "tester";
|
||||
char ky[] = "key";
|
||||
xce_peerinfo_t from = { .name = nm, .key = ky };
|
||||
|
||||
TEST_ASSERT_EQUAL_INT(-EINVAL,
|
||||
xce_job_state_query_req(NULL, &from, &query, NULL, NULL, 0));
|
||||
|
||||
endpoint_t* ep = new_test_endpoint();
|
||||
TEST_ASSERT_EQUAL_INT(-EINVAL,
|
||||
xce_job_state_query_req(ep, NULL, &query, NULL, NULL, 0));
|
||||
TEST_ASSERT_EQUAL_INT(-EINVAL,
|
||||
xce_job_state_query_req(ep, &from, NULL, NULL, NULL, 0));
|
||||
|
||||
endpoint_fini(ep);
|
||||
free(ep);
|
||||
printf("test_job_state_query_req_null_args: PASS\n");
|
||||
}
|
||||
|
||||
static void test_job_state_query_roundtrip_normal(void)
|
||||
{
|
||||
uint64_t jobids[5] = { 101, 202, 303, 404, 505 };
|
||||
xce_job_state_query_t query = {
|
||||
.n_total = 5, .pos = 0, .count = 5, .jobids = jobids
|
||||
};
|
||||
char nm[] = "tester";
|
||||
char ky[] = "key";
|
||||
xce_peerinfo_t from = { .name = nm, .key = ky };
|
||||
|
||||
endpoint_t* ep = new_test_endpoint();
|
||||
|
||||
/* pktlimit <= 0 -> whole list in one page. */
|
||||
TEST_ASSERT_EQUAL_INT(0,
|
||||
xce_job_state_query_req(ep, &from, &query, NULL, NULL, 0));
|
||||
|
||||
xce_request_t decoded;
|
||||
pop_and_decode(ep, &decoded);
|
||||
|
||||
TEST_ASSERT_EQUAL_INT(OPCode_jobStateQuery, decoded.opcode);
|
||||
TEST_ASSERT_EQUAL_INT(Request_body_jobstatequerymsg, decoded.body_which);
|
||||
TEST_ASSERT_NOT_NULL(decoded.body.jobstatequerymsg);
|
||||
TEST_ASSERT_EQUAL_INT(5, decoded.body.jobstatequerymsg->n_total);
|
||||
TEST_ASSERT_EQUAL_INT(0, decoded.body.jobstatequerymsg->pos);
|
||||
TEST_ASSERT_EQUAL_INT(5, decoded.body.jobstatequerymsg->count);
|
||||
TEST_ASSERT_NOT_NULL(decoded.body.jobstatequerymsg->jobids);
|
||||
for (int i = 0; i < 5; i++) {
|
||||
TEST_ASSERT_EQUAL_UINT64(jobids[i], decoded.body.jobstatequerymsg->jobids[i]);
|
||||
}
|
||||
|
||||
xce_free_request(&decoded);
|
||||
assert_queue_empty(ep);
|
||||
|
||||
endpoint_fini(ep);
|
||||
free(ep);
|
||||
printf("test_job_state_query_roundtrip_normal: PASS\n");
|
||||
}
|
||||
|
||||
static void test_job_state_query_roundtrip_empty(void)
|
||||
{
|
||||
xce_job_state_query_t query = {
|
||||
.n_total = 0, .pos = 0, .count = 0, .jobids = NULL
|
||||
};
|
||||
char nm[] = "tester";
|
||||
char ky[] = "key";
|
||||
xce_peerinfo_t from = { .name = nm, .key = ky };
|
||||
|
||||
endpoint_t* ep = new_test_endpoint();
|
||||
|
||||
TEST_ASSERT_EQUAL_INT(0,
|
||||
xce_job_state_query_req(ep, &from, &query, NULL, NULL, 0));
|
||||
|
||||
/* Req 8.1: zero jobs to query is still a complete, single-page exchange
|
||||
* -- not skipped, not left unsent. */
|
||||
xce_request_t decoded;
|
||||
pop_and_decode(ep, &decoded);
|
||||
|
||||
TEST_ASSERT_EQUAL_INT(OPCode_jobStateQuery, decoded.opcode);
|
||||
TEST_ASSERT_EQUAL_INT(Request_body_jobstatequerymsg, decoded.body_which);
|
||||
TEST_ASSERT_NOT_NULL(decoded.body.jobstatequerymsg);
|
||||
TEST_ASSERT_EQUAL_INT(0, decoded.body.jobstatequerymsg->n_total);
|
||||
TEST_ASSERT_EQUAL_INT(0, decoded.body.jobstatequerymsg->pos);
|
||||
TEST_ASSERT_EQUAL_INT(0, decoded.body.jobstatequerymsg->count);
|
||||
TEST_ASSERT_NULL(decoded.body.jobstatequerymsg->jobids);
|
||||
|
||||
xce_free_request(&decoded);
|
||||
assert_queue_empty(ep);
|
||||
|
||||
endpoint_fini(ep);
|
||||
free(ep);
|
||||
printf("test_job_state_query_roundtrip_empty: PASS\n");
|
||||
}
|
||||
|
||||
static void test_job_state_query_roundtrip_multipage(void)
|
||||
{
|
||||
uint64_t jobids[7] = { 1, 2, 3, 4, 5, 6, 7 };
|
||||
xce_job_state_query_t query = {
|
||||
.n_total = 7, .pos = 0, .count = 7, .jobids = jobids
|
||||
};
|
||||
char nm[] = "tester";
|
||||
char ky[] = "key";
|
||||
xce_peerinfo_t from = { .name = nm, .key = ky };
|
||||
|
||||
endpoint_t* ep = new_test_endpoint();
|
||||
|
||||
/* pktlimit=3 over 7 jobids -> pages of 3, 3, 1 (Req 8.1/8.3: nothing
|
||||
* truncated or dropped across pages). */
|
||||
TEST_ASSERT_EQUAL_INT(0,
|
||||
xce_job_state_query_req(ep, &from, &query, NULL, NULL, 3));
|
||||
|
||||
uint64_t rebuilt[7];
|
||||
int rebuilt_n = 0;
|
||||
int expected_counts[3] = { 3, 3, 1 };
|
||||
int expected_pos[3] = { 0, 3, 6 };
|
||||
|
||||
for (int page = 0; page < 3; page++) {
|
||||
xce_request_t decoded;
|
||||
pop_and_decode(ep, &decoded);
|
||||
|
||||
TEST_ASSERT_EQUAL_INT(OPCode_jobStateQuery, decoded.opcode);
|
||||
TEST_ASSERT_EQUAL_INT(Request_body_jobstatequerymsg, decoded.body_which);
|
||||
TEST_ASSERT_NOT_NULL(decoded.body.jobstatequerymsg);
|
||||
TEST_ASSERT_EQUAL_INT(7, decoded.body.jobstatequerymsg->n_total);
|
||||
TEST_ASSERT_EQUAL_INT(expected_pos[page], decoded.body.jobstatequerymsg->pos);
|
||||
TEST_ASSERT_EQUAL_INT(expected_counts[page], decoded.body.jobstatequerymsg->count);
|
||||
|
||||
for (int i = 0; i < decoded.body.jobstatequerymsg->count; i++) {
|
||||
rebuilt[rebuilt_n++] = decoded.body.jobstatequerymsg->jobids[i];
|
||||
}
|
||||
|
||||
xce_free_request(&decoded);
|
||||
}
|
||||
|
||||
assert_queue_empty(ep);
|
||||
|
||||
TEST_ASSERT_EQUAL_INT(7, rebuilt_n);
|
||||
for (int i = 0; i < 7; i++) {
|
||||
TEST_ASSERT_EQUAL_UINT64(jobids[i], rebuilt[i]);
|
||||
}
|
||||
|
||||
endpoint_fini(ep);
|
||||
free(ep);
|
||||
printf("test_job_state_query_roundtrip_multipage: PASS\n");
|
||||
}
|
||||
|
||||
/* ── OPCode_jobStateReport round trips (worker -> controller) ───────────── */
|
||||
|
||||
static void test_job_state_report_req_null_args(void)
|
||||
{
|
||||
xce_job_state_report_t report = { 0 };
|
||||
char nm[] = "tester";
|
||||
char ky[] = "key";
|
||||
xce_peerinfo_t from = { .name = nm, .key = ky };
|
||||
|
||||
TEST_ASSERT_EQUAL_INT(-EINVAL,
|
||||
xce_job_state_report_req(NULL, &from, &report, NULL, NULL, 0));
|
||||
|
||||
endpoint_t* ep = new_test_endpoint();
|
||||
TEST_ASSERT_EQUAL_INT(-EINVAL,
|
||||
xce_job_state_report_req(ep, NULL, &report, NULL, NULL, 0));
|
||||
TEST_ASSERT_EQUAL_INT(-EINVAL,
|
||||
xce_job_state_report_req(ep, &from, NULL, NULL, NULL, 0));
|
||||
|
||||
endpoint_fini(ep);
|
||||
free(ep);
|
||||
printf("test_job_state_report_req_null_args: PASS\n");
|
||||
}
|
||||
|
||||
/* Covers all four JobStateStatus values and both queried true/false, plus
|
||||
* exitCode/finishTimeMs survival for finished entries. */
|
||||
static void test_job_state_report_roundtrip_normal(void)
|
||||
{
|
||||
xce_job_state_entry_t e0 = {
|
||||
.jobid = 9001, .status = JobStateStatus_running,
|
||||
.exit_code = 0, .finish_time_ms = 0, .queried = 1
|
||||
};
|
||||
xce_job_state_entry_t e1 = {
|
||||
.jobid = 9002, .status = JobStateStatus_finishedOk,
|
||||
.exit_code = 0, .finish_time_ms = 1693500000123ULL, .queried = 0
|
||||
};
|
||||
xce_job_state_entry_t e2 = {
|
||||
.jobid = 9003, .status = JobStateStatus_finishedFail,
|
||||
.exit_code = 137, .finish_time_ms = 1693500005555ULL, .queried = 1
|
||||
};
|
||||
xce_job_state_entry_t e3 = {
|
||||
.jobid = 9004, .status = JobStateStatus_unrecognized,
|
||||
.exit_code = 0, .finish_time_ms = 0, .queried = 0
|
||||
};
|
||||
xce_job_state_entry_t* entries[4] = { &e0, &e1, &e2, &e3 };
|
||||
xce_job_state_report_t report = {
|
||||
.n_total = 4, .pos = 0, .count = 4, .entries = entries
|
||||
};
|
||||
char nm[] = "tester";
|
||||
char ky[] = "key";
|
||||
xce_peerinfo_t from = { .name = nm, .key = ky };
|
||||
|
||||
endpoint_t* ep = new_test_endpoint();
|
||||
|
||||
TEST_ASSERT_EQUAL_INT(0,
|
||||
xce_job_state_report_req(ep, &from, &report, NULL, NULL, 0));
|
||||
|
||||
xce_request_t decoded;
|
||||
pop_and_decode(ep, &decoded);
|
||||
|
||||
TEST_ASSERT_EQUAL_INT(OPCode_jobStateReport, decoded.opcode);
|
||||
TEST_ASSERT_EQUAL_INT(Request_body_jobstatereportmsg, decoded.body_which);
|
||||
TEST_ASSERT_NOT_NULL(decoded.body.jobstatereportmsg);
|
||||
TEST_ASSERT_EQUAL_INT(4, decoded.body.jobstatereportmsg->n_total);
|
||||
TEST_ASSERT_EQUAL_INT(0, decoded.body.jobstatereportmsg->pos);
|
||||
TEST_ASSERT_EQUAL_INT(4, decoded.body.jobstatereportmsg->count);
|
||||
TEST_ASSERT_NOT_NULL(decoded.body.jobstatereportmsg->entries);
|
||||
|
||||
xce_job_state_entry_t** got = decoded.body.jobstatereportmsg->entries;
|
||||
|
||||
TEST_ASSERT_EQUAL_UINT64(9001, got[0]->jobid);
|
||||
TEST_ASSERT_EQUAL_INT(JobStateStatus_running, got[0]->status);
|
||||
TEST_ASSERT_EQUAL_INT(1, got[0]->queried);
|
||||
|
||||
TEST_ASSERT_EQUAL_UINT64(9002, got[1]->jobid);
|
||||
TEST_ASSERT_EQUAL_INT(JobStateStatus_finishedOk, got[1]->status);
|
||||
TEST_ASSERT_EQUAL_INT(0, got[1]->queried);
|
||||
TEST_ASSERT_EQUAL_UINT64(1693500000123ULL, got[1]->finish_time_ms);
|
||||
|
||||
TEST_ASSERT_EQUAL_UINT64(9003, got[2]->jobid);
|
||||
TEST_ASSERT_EQUAL_INT(JobStateStatus_finishedFail, got[2]->status);
|
||||
TEST_ASSERT_EQUAL_INT(1, got[2]->queried);
|
||||
TEST_ASSERT_EQUAL_INT32(137, got[2]->exit_code);
|
||||
TEST_ASSERT_EQUAL_UINT64(1693500005555ULL, got[2]->finish_time_ms);
|
||||
|
||||
TEST_ASSERT_EQUAL_UINT64(9004, got[3]->jobid);
|
||||
TEST_ASSERT_EQUAL_INT(JobStateStatus_unrecognized, got[3]->status);
|
||||
TEST_ASSERT_EQUAL_INT(0, got[3]->queried);
|
||||
|
||||
xce_free_request(&decoded);
|
||||
assert_queue_empty(ep);
|
||||
|
||||
endpoint_fini(ep);
|
||||
free(ep);
|
||||
printf("test_job_state_report_roundtrip_normal: PASS\n");
|
||||
}
|
||||
|
||||
static void test_job_state_report_roundtrip_empty(void)
|
||||
{
|
||||
xce_job_state_report_t report = {
|
||||
.n_total = 0, .pos = 0, .count = 0, .entries = NULL
|
||||
};
|
||||
char nm[] = "tester";
|
||||
char ky[] = "key";
|
||||
xce_peerinfo_t from = { .name = nm, .key = ky };
|
||||
|
||||
endpoint_t* ep = new_test_endpoint();
|
||||
|
||||
TEST_ASSERT_EQUAL_INT(0,
|
||||
xce_job_state_report_req(ep, &from, &report, NULL, NULL, 0));
|
||||
|
||||
xce_request_t decoded;
|
||||
pop_and_decode(ep, &decoded);
|
||||
|
||||
TEST_ASSERT_EQUAL_INT(OPCode_jobStateReport, decoded.opcode);
|
||||
TEST_ASSERT_EQUAL_INT(Request_body_jobstatereportmsg, decoded.body_which);
|
||||
TEST_ASSERT_NOT_NULL(decoded.body.jobstatereportmsg);
|
||||
TEST_ASSERT_EQUAL_INT(0, decoded.body.jobstatereportmsg->n_total);
|
||||
TEST_ASSERT_EQUAL_INT(0, decoded.body.jobstatereportmsg->pos);
|
||||
TEST_ASSERT_EQUAL_INT(0, decoded.body.jobstatereportmsg->count);
|
||||
TEST_ASSERT_NULL(decoded.body.jobstatereportmsg->entries);
|
||||
|
||||
xce_free_request(&decoded);
|
||||
assert_queue_empty(ep);
|
||||
|
||||
endpoint_fini(ep);
|
||||
free(ep);
|
||||
printf("test_job_state_report_roundtrip_empty: PASS\n");
|
||||
}
|
||||
|
||||
static void test_job_state_report_roundtrip_multipage(void)
|
||||
{
|
||||
xce_job_state_entry_t e[5];
|
||||
xce_job_state_entry_t* entries[5];
|
||||
for (int i = 0; i < 5; i++) {
|
||||
memset(&e[i], 0, sizeof(e[i]));
|
||||
e[i].jobid = 8000 + (uint64_t)i;
|
||||
e[i].status = JobStateStatus_finishedOk;
|
||||
e[i].finish_time_ms = 1700000000000ULL + (uint64_t)i;
|
||||
e[i].queried = (i % 2 == 0) ? 1 : 0;
|
||||
entries[i] = &e[i];
|
||||
}
|
||||
xce_job_state_report_t report = {
|
||||
.n_total = 5, .pos = 0, .count = 5, .entries = entries
|
||||
};
|
||||
char nm[] = "tester";
|
||||
char ky[] = "key";
|
||||
xce_peerinfo_t from = { .name = nm, .key = ky };
|
||||
|
||||
endpoint_t* ep = new_test_endpoint();
|
||||
|
||||
/* pktlimit=2 over 5 entries -> pages of 2, 2, 1. */
|
||||
TEST_ASSERT_EQUAL_INT(0,
|
||||
xce_job_state_report_req(ep, &from, &report, NULL, NULL, 2));
|
||||
|
||||
uint64_t rebuilt_jobids[5];
|
||||
int rebuilt_queried[5];
|
||||
int rebuilt_n = 0;
|
||||
int expected_counts[3] = { 2, 2, 1 };
|
||||
int expected_pos[3] = { 0, 2, 4 };
|
||||
|
||||
for (int page = 0; page < 3; page++) {
|
||||
xce_request_t decoded;
|
||||
pop_and_decode(ep, &decoded);
|
||||
|
||||
TEST_ASSERT_EQUAL_INT(OPCode_jobStateReport, decoded.opcode);
|
||||
TEST_ASSERT_EQUAL_INT(Request_body_jobstatereportmsg, decoded.body_which);
|
||||
TEST_ASSERT_NOT_NULL(decoded.body.jobstatereportmsg);
|
||||
TEST_ASSERT_EQUAL_INT(5, decoded.body.jobstatereportmsg->n_total);
|
||||
TEST_ASSERT_EQUAL_INT(expected_pos[page], decoded.body.jobstatereportmsg->pos);
|
||||
TEST_ASSERT_EQUAL_INT(expected_counts[page], decoded.body.jobstatereportmsg->count);
|
||||
|
||||
for (int i = 0; i < decoded.body.jobstatereportmsg->count; i++) {
|
||||
xce_job_state_entry_t* got = decoded.body.jobstatereportmsg->entries[i];
|
||||
rebuilt_jobids[rebuilt_n] = got->jobid;
|
||||
rebuilt_queried[rebuilt_n] = got->queried;
|
||||
rebuilt_n++;
|
||||
}
|
||||
|
||||
xce_free_request(&decoded);
|
||||
}
|
||||
|
||||
assert_queue_empty(ep);
|
||||
|
||||
TEST_ASSERT_EQUAL_INT(5, rebuilt_n);
|
||||
for (int i = 0; i < 5; i++) {
|
||||
TEST_ASSERT_EQUAL_UINT64(e[i].jobid, rebuilt_jobids[i]);
|
||||
TEST_ASSERT_EQUAL_INT(e[i].queried, rebuilt_queried[i]);
|
||||
}
|
||||
|
||||
endpoint_fini(ep);
|
||||
free(ep);
|
||||
printf("test_job_state_report_roundtrip_multipage: PASS\n");
|
||||
}
|
||||
|
||||
void test_job_state_wire(void)
|
||||
{
|
||||
printf("\n>> job-state reconciliation wire codec round-trip tests\n\n");
|
||||
RUN_TEST(test_job_state_query_req_null_args);
|
||||
RUN_TEST(test_job_state_query_roundtrip_normal);
|
||||
RUN_TEST(test_job_state_query_roundtrip_empty);
|
||||
RUN_TEST(test_job_state_query_roundtrip_multipage);
|
||||
RUN_TEST(test_job_state_report_req_null_args);
|
||||
RUN_TEST(test_job_state_report_roundtrip_normal);
|
||||
RUN_TEST(test_job_state_report_roundtrip_empty);
|
||||
RUN_TEST(test_job_state_report_roundtrip_multipage);
|
||||
}
|
||||
Loading…
Add table
Add a link
Reference in a new issue