DAOS-19616 placement: derive CURRENT layouts from PRE and POST

[liuxuezhao]

Generating CURRENT directly can mis-handle cascading shard relocation: a shard which an earlier rebuild put on a spare can be handed a different spare by a new, unrelated failure, so the CURRENT layout names a target that was never written to while the PRE_REBUILD layout, used by the migration fetch and every read-only client, keeps reading the old one.

CURRENT is now a transition view derived from the two canonical endpoints instead of an independent generation mode:

• Build POST_REBUILD first; if its placement met no transitional target (DOWN/DRAIN/UP/NEW), POST is CURRENT (single-layout path).
• Otherwise generate PRE_REBUILD and derive CURRENT from both while preserving every readable PRE target as the primary.
• Add the POST target as a write-only peer for drain, reintegration, extension and collateral relocation of a healthy shard.
• Use POST directly for failed PRE targets and unaffected objects.
• Handle missing rebuild spares without invalid target lookups. The CURRENT branches of comp_need_remap() and the separate extend pass are removed; readable targets equal PRE, write destinations equal POST, so the placement, migration and rebuild-scan views can no longer disagree.

Placement UTs: PLACEMENT 18 now rejects readable duplicates only; a target can legitimately be the readable primary of one shard and the rebuild destination of another during extension plus reintegration. Add no_stale_read_source, extension_preserves_read_sources, no_hidden_peer_source and relocation_peer_alloc_failure, asserting that every healthy PRE read source stays in the CURRENT write set, that every readable CURRENT target was written before the failure and that pl_map_extend() failures are propagated. Test pool maps give targets fseq 1 as gen_pool_buf() does.

Also refine the migration comment on a fetch returning iod_size 0 (placement mismatch, older fetch epoch, unresolved DTX) and why -DER_DATA_LOSS is excluded from the rebuild status.

Build and run the reproducer from the repository root:

scons --config=force --jobs=16
install/bin/jump_pl_map install/lib64/libdaos.so
./utils/run_utest.py --suite_filter='^placement$' --sudo=no

Without the placement change the new tests fail on a PRE_REBUILD read source that CURRENT does not write to; with it all 43 cases pass.

Features: rebuild

Steps for the author:

• Commit message follows the guidelines.
• Appropriate Features or Test-tag pragmas were used.
• Appropriate Functional Test Stages were run.
• At least two positive code reviews including at least one code owner from each category referenced in the PR.
• Testing is complete. If necessary, forced-landing label added and a reason added in a comment.

After all prior steps are complete:

• Gatekeeper requested (daos-gatekeeper added as a reviewer).

[github-actions]

Ticket title is 'FIO crc mismatch detected on 4x node IB config while running HSN cable pulls test'
Status is 'Open'
Job should run at elevated priority (1)
https://daosio.atlassian.net/browse/DAOS-19616