https://hesela.com/New analyses by their recorded publication date.enhttps://hesela.com/analyses/emmc-ufs-rpmb-boundaries/https://hesela.com/analyses/emmc-ufs-rpmb-boundaries/#publication-emmc-ufs-rpmb-boundariesSat, 17 Oct 2026 00:00:00 GMTHow managed flash differs from raw NAND, why eMMC and UFS are not interchangeable interfaces, and why RPMB authentication is neither encryption nor a backup. https://hesela.com/analyses/iscsi-identity-sessions-recovery/https://hesela.com/analyses/iscsi-identity-sessions-recovery/#publication-iscsi-identity-sessions-recoverySat, 17 Oct 2026 00:00:00 GMTSeparate iSCSI names, portals, sessions, command ordering and recovery levels, with a source-checked distinction between projected and measured performance. https://hesela.com/analyses/magnetic-retention-time-and-probability/https://hesela.com/analyses/magnetic-retention-time-and-probability/#publication-magnetic-retention-time-and-probabilitySat, 17 Oct 2026 00:00:00 GMTA reproducible two-state model separates first reversal, final-state error and magnetization decay, and shows why a mean energy barrier cannot predict retention. https://hesela.com/analyses/mmap-visibility-size-durability/https://hesela.com/analyses/mmap-visibility-size-durability/#publication-mmap-visibility-size-durabilitySat, 17 Oct 2026 00:00:00 GMTSeven reproducible Linux checks distinguish mapped visibility, file size, partial-page tails, SIGBUS and sync argument errors from power-loss durability. https://hesela.com/analyses/nvme-fabrics-transports-discovery/https://hesela.com/analyses/nvme-fabrics-transports-discovery/#publication-nvme-fabrics-transports-discoverySat, 17 Oct 2026 00:00:00 GMTDistinguish NVMe/TCP, NVMe/RDMA and NVMe/FC from discovery, path selection and security, without treating a transport label as a performance guarantee. https://hesela.com/analyses/phi-failure-detectors-suspicion/https://hesela.com/analyses/phi-failure-detectors-suspicion/#publication-phi-failure-detectors-suspicionSat, 17 Oct 2026 00:00:00 GMTWhat a phi score measures, why Cassandra's score differs from the original model, and why suspecting a replica does not fence it. Includes reproducible synthetic examples. https://hesela.com/analyses/sieve-cache-eviction-traces/https://hesela.com/analyses/sieve-cache-eviction-traces/#publication-sieve-cache-eviction-tracesSat, 17 Oct 2026 00:00:00 GMTExplain SIEVE versus CLOCK and LRU, replay cache traces and FIFO capacity anomalies, and separate miss counts from storage performance claims. https://hesela.com/analyses/sleeping-disks-partial-reconstruction/https://hesela.com/analyses/sleeping-disks-partial-reconstruction/#publication-sleeping-disks-partial-reconstructionSat, 17 Oct 2026 00:00:00 GMTAn exhaustive check of two historical XOR codes separates serving a read without spin-up from recovering every data symbol. Includes code, all 512 availability states and explicit limits. https://hesela.com/analyses/tmpfile-publication-rename-contracts/https://hesela.com/analyses/tmpfile-publication-rename-contracts/#publication-tmpfile-publication-rename-contractsSat, 17 Oct 2026 00:00:00 GMTA native Linux observation separates O_TMPFILE staging, linkat publication, no-replace rename and atomic exchange from crash durability. https://hesela.com/analyses/b-epsilon-trees-buffered-updates/https://hesela.com/analyses/b-epsilon-trees-buffered-updates/#publication-b-epsilon-trees-buffered-updatesFri, 16 Oct 2026 00:00:00 GMTHow buffered tree updates remain query-visible, why blind upserts differ from read-modify-write, and why equal flush counts can hide different completion times. https://hesela.com/analyses/hdd-adjacent-track-interference-refresh/https://hesela.com/analyses/hdd-adjacent-track-interference-refresh/#publication-hdd-adjacent-track-interference-refreshFri, 16 Oct 2026 00:00:00 GMTRepeated HDD writes can trigger protective refreshes even with CMR. Separate write-induced ATI, read-channel interference and SMR band rewrites before diagnosing latency. https://hesela.com/analyses/hdd-annualized-workload-rate/https://hesela.com/analyses/hdd-annualized-workload-rate/#publication-hdd-annualized-workload-rateFri, 16 Oct 2026 00:00:00 GMTAn HDD workload rating is not SSD TBW or a failure countdown. Power-on hours, read traffic and maintenance scans change what the annualized number means. https://hesela.com/analyses/hdd-write-verification-iraw/https://hesela.com/analyses/hdd-write-verification-iraw/#publication-hdd-write-verification-irawFri, 16 Oct 2026 00:00:00 GMTRead-after-write verification, IRAW and SCSI VERIFY answer different questions from flush or FUA. Learn what a successful check does and does not establish. https://hesela.com/analyses/lmdb-readers-page-reuse-map-full/https://hesela.com/analyses/lmdb-readers-page-reuse-map-full/#publication-lmdb-readers-page-reuse-map-fullFri, 16 Oct 2026 00:00:00 GMTA controlled LMDB experiment separates reader lifetime, reusable pages and map capacity: why a live reader can stop writes without holding the writer mutex. https://hesela.com/analyses/nfs-silly-rename-open-unlink/https://hesela.com/analyses/nfs-silly-rename-open-unlink/#publication-nfs-silly-rename-open-unlinkFri, 16 Oct 2026 00:00:00 GMTSilly rename, open-file preservation and a misleading Linux comment: trace NFS deletion from the 1985 paper to a pinned implementation and RFC erratum. https://hesela.com/analyses/raidz-expansion-parity-capacity/https://hesela.com/analyses/raidz-expansion-parity-capacity/#publication-raidz-expansion-parity-capacityFri, 16 Oct 2026 00:00:00 GMTReflow spreads old blocks across more disks without changing their data-to-parity ratio. A reproducible model separates retained parity overhead, reported space and snapshot-aware rewriting. https://hesela.com/analyses/btrfs-tree-log-read-only-replay/https://hesela.com/analyses/btrfs-tree-log-read-only-replay/#publication-btrfs-tree-log-read-only-replayThu, 15 Oct 2026 00:00:00 GMTSeparate Btrfs fsync logging, transaction commit, read-only mount recovery and destructive log clearing, with a reproducible Linux 6.15/6.16 option audit. https://hesela.com/analyses/f2fs-checkpoint-fsync-recovery/https://hesela.com/analyses/f2fs-checkpoint-fsync-recovery/#publication-f2fs-checkpoint-fsync-recoveryThu, 15 Oct 2026 00:00:00 GMTA pinned Linux source audit separates F2FS checkpoints, roll-forward recovery and the fsync success path when checkpointing is disabled. https://hesela.com/analyses/file-clones-deferred-copy-cost/https://hesela.com/analyses/file-clones-deferred-copy-cost/#publication-file-clones-deferred-copy-costThu, 15 Oct 2026 00:00:00 GMTSeparate reflinks, hard links and in-kernel copies; trace deferred costs, Btrfs maintenance, shared-space accounting and FAST 2020's research limits. https://hesela.com/analyses/nvme-multipath-ana-path-selection/https://hesela.com/analyses/nvme-multipath-ana-path-selection/#publication-nvme-multipath-ana-path-selectionThu, 15 Oct 2026 00:00:00 GMTSeparate NVMe namespace identity from ANA access states: NSID scope, identifier lifetime, Linux path policies and controller-level queue accounting. https://hesela.com/analyses/optical-disc-lifetime-estimates/https://hesela.com/analyses/optical-disc-lifetime-estimates/#publication-optical-disc-lifetime-estimatesThu, 15 Oct 2026 00:00:00 GMTOptical-disc lifetime is not a warranty: CD/DVD error units, survival estimates, Blu-ray HTL/LTH, and a source-level check of M-DISC longevity claims. https://hesela.com/analyses/soft-updates-metadata-ordering/https://hesela.com/analyses/soft-updates-metadata-ordering/#publication-soft-updates-metadata-orderingThu, 15 Oct 2026 00:00:00 GMTWhy write ordering can leave reclaimable space instead of unsafe pointers, and why background fsck and metadata consistency do not prove application durability. https://hesela.com/analyses/xfs-delayed-logging-commit-boundaries/https://hesela.com/analyses/xfs-delayed-logging-commit-boundaries/#publication-xfs-delayed-logging-commit-boundariesThu, 15 Oct 2026 00:00:00 GMTSeparate XFS commit from persistence, relogging from deferred recovery, and journal atomicity from online visibility. https://hesela.com/analyses/ext4-journal-replay-revocation/https://hesela.com/analyses/ext4-journal-replay-revocation/#publication-ext4-journal-replay-revocationWed, 14 Oct 2026 00:00:00 GMTJBD2 replay, checksum failure paths and ambiguous log tails, with source-checked recovery boundaries and reproducible counterexamples. https://hesela.com/analyses/filesystem-aging-free-space-fragmentation/https://hesela.com/analyses/filesystem-aging-free-space-fragmentation/#publication-filesystem-aging-free-space-fragmentationWed, 14 Oct 2026 00:00:00 GMTSeparate filesystem layout history, free-space fragmentation and SSD conditioning, with three USENIX studies and a reproducible allocation counterexample. https://hesela.com/analyses/io-uring-completion-ordering-durability/https://hesela.com/analyses/io-uring-completion-ordering-durability/#publication-io-uring-completion-ordering-durabilityWed, 14 Oct 2026 00:00:00 GMTio_uring write ordering, protection information, cancellation CQEs and linked timeouts: what each result proves, and what still needs synchronization. https://hesela.com/analyses/rocksdb-wal-flush-manifest-boundaries/https://hesela.com/analyses/rocksdb-wal-flush-manifest-boundaries/#publication-rocksdb-wal-flush-manifest-boundariesWed, 14 Oct 2026 00:00:00 GMTRocksDB WAL sync, memtable flush, write stalls, read snapshots, checkpoints and backups have distinct recovery and retention contracts. https://hesela.com/analyses/smb-handles-leases-durability/https://hesela.com/analyses/smb-handles-leases-durability/#publication-smb-handles-leases-durabilityWed, 14 Oct 2026 00:00:00 GMTDurable and persistent SMB handles, caching leases and write-through protect different boundaries. Learn what to verify after a disconnect or failover. https://hesela.com/analyses/sqlite-wal-checkpoint-starvation/https://hesela.com/analyses/sqlite-wal-checkpoint-starvation/#publication-sqlite-wal-checkpoint-starvationWed, 14 Oct 2026 00:00:00 GMTMeasured SQLite WAL checkpoints and stale-snapshot retries: why restarting a transaction can still overwrite an update when the application reuses old values. https://hesela.com/analyses/zfs-zil-slog-durability-boundaries/https://hesela.com/analyses/zfs-zil-slog-durability-boundaries/#publication-zfs-zil-slog-durability-boundariesWed, 14 Oct 2026 00:00:00 GMTZIL, SLOG and transaction groups protect different boundaries. Pinned OpenZFS code explains indirect records, sync settings and misleading log-byte counters. https://hesela.com/analyses/backup-verification-is-not-a-restore-test/https://hesela.com/analyses/backup-verification-is-not-a-restore-test/#publication-backup-verification-is-not-a-restore-testTue, 13 Oct 2026 00:00:00 GMTSeparate backup completion, integrity, freshness and usable recovery, with reproducible SQLite and rsync experiments and explicit limits. https://hesela.com/analyses/o-direct-alignment-by-kernel/https://hesela.com/analyses/o-direct-alignment-by-kernel/#publication-o-direct-alignment-by-kernelTue, 13 Oct 2026 00:00:00 GMTQuery O_DIRECT per-file alignment and distinguish cache hints, prefetch, NOWAIT and mincore residency from actual I/O and durable writes. https://hesela.com/analyses/parity-declustering-rebuild-load/https://hesela.com/analyses/parity-declustering-rebuild-load/#publication-parity-declustering-rebuild-loadTue, 13 Oct 2026 00:00:00 GMTAn executable eight-disk example separates repair-read volume, survivor load and stripe width, with evidence from ASPLOS, Panasas and OpenZFS dRAID. https://hesela.com/analyses/pmem-persistence-domains-dax/https://hesela.com/analyses/pmem-persistence-domains-dax/#publication-pmem-persistence-domains-daxTue, 13 Oct 2026 00:00:00 GMTADR, eADR and MAP_SYNC protect different boundaries. A source-linked review and executable crash-state example separate addressability, visibility, persistence and atomicity. https://hesela.com/analyses/reed-solomon-matrix-recovery-guarantees/https://hesela.com/analyses/reed-solomon-matrix-recovery-guarantees/#publication-reed-solomon-matrix-recovery-guaranteesTue, 13 Oct 2026 00:00:00 GMTExact GF(7) examples separate MDS recovery, missing fragments and silent corruption, showing why even a unique decoded answer can be wrong. https://hesela.com/analyses/replica-repair-coverage/https://hesela.com/analyses/replica-repair-coverage/#publication-replica-repair-coverageTue, 13 Oct 2026 00:00:00 GMTReplica repair coverage, quorum overlap and set reconciliation: why matching keys, successful responses and compact summaries do not prove identical durable state. https://hesela.com/analyses/snapshot-backup-recovery-boundaries/https://hesela.com/analyses/snapshot-backup-recovery-boundaries/#publication-snapshot-backup-recovery-boundariesTue, 13 Oct 2026 00:00:00 GMTSeparate snapshot consistency, independent recovery copies and RPO/RTO targets, with source-backed OpenZFS, PostgreSQL and VSS examples. https://hesela.com/analyses/sparse-files-preallocation-thin-capacity/https://hesela.com/analyses/sparse-files-preallocation-thin-capacity/#publication-sparse-files-preallocation-thin-capacityTue, 13 Oct 2026 00:00:00 GMTFile size, allocated blocks and thin-pool headroom are different measurements. Learn what preallocation, hole punching and zero-filled reads actually establish. https://hesela.com/analyses/bloom-filters-lsm-read-cost/https://hesela.com/analyses/bloom-filters-lsm-read-cost/#publication-bloom-filters-lsm-read-costMon, 12 Oct 2026 00:00:00 GMTLSM lookup costs, filter retention and counting limits, with reproducible allocation, deletion and memory examples grounded in primary sources. https://hesela.com/analyses/content-defined-chunking-boundaries/https://hesela.com/analyses/content-defined-chunking-boundaries/#publication-content-defined-chunking-boundariesMon, 12 Oct 2026 00:00:00 GMTReproducible FastCDC examples show how file framing, parameter changes and terminal chunks alter measured reuse, even when file contents remain unchanged. https://hesela.com/analyses/crash-testing-oracles-and-bounds/https://hesela.com/analyses/crash-testing-oracles-and-bounds/#publication-crash-testing-oracles-and-boundsMon, 12 Oct 2026 00:00:00 GMTWhat ALICE, CrashMonkey/ACE, SQLite and Ananke actually test: fault boundaries, separate recovery outcomes and an executable correctness-oracle example. https://hesela.com/analyses/key-value-separation-cleaning-cost/https://hesela.com/analyses/key-value-separation-cleaning-cost/#publication-key-value-separation-cleaning-costMon, 12 Oct 2026 00:00:00 GMTWiscKey, HashKV and BlobDB separate value movement from key sorting. A reproducible byte model shows when value-log cleaning can erase the saving. https://hesela.com/analyses/linux-atomic-writes-durability/https://hesela.com/analyses/linux-atomic-writes-durability/#publication-linux-atomic-writes-durabilityMon, 12 Oct 2026 00:00:00 GMTAtomic writes, durable completion and powersafe overwrite protect different boundaries. Check Linux request geometry and SQLite's outside-range assumption. https://hesela.com/analyses/lsm-tombstones-delete-persistence/https://hesela.com/analyses/lsm-tombstones-delete-persistence/#publication-lsm-tombstones-delete-persistenceMon, 12 Oct 2026 00:00:00 GMTSeparate delete visibility, crash durability and space reclamation: snapshots, Lethe's delete-aware compaction and RocksDB's range-tombstone safeguards. https://hesela.com/analyses/nfs-stable-writes-flush-fua/https://hesela.com/analyses/nfs-stable-writes-flush-fua/#publication-nfs-stable-writes-flush-fuaMon, 12 Oct 2026 00:00:00 GMTNFSv3 stability, COMMIT verifiers and Linux flush/FUA: distinguish durable writes from cache visibility, with bounded NFSv4 delegation comparisons. https://hesela.com/analyses/nvme-fdp-placement-and-limits/https://hesela.com/analyses/nvme-fdp-placement-and-limits/#publication-nvme-fdp-placement-and-limitsMon, 12 Oct 2026 00:00:00 GMTHow reclaim units, handles and placement identifiers fit together, why FDP differs from ZNS, and what a credible write-amplification comparison must measure. https://hesela.com/analyses/object-etags-checksums-conditional-writes/https://hesela.com/analyses/object-etags-checksums-conditional-writes/#publication-object-etags-checksums-conditional-writesMon, 12 Oct 2026 00:00:00 GMTS3 ETags, full-object and composite checksums, version IDs and write preconditions: what each validates, and how to avoid checking download metadata instead of bytes. https://hesela.com/analyses/optical-disc-rot-error-rate-signals/https://hesela.com/analyses/optical-disc-rot-error-rate-signals/#publication-optical-disc-rot-error-rate-signalsMon, 12 Oct 2026 00:00:00 GMTRecordable discs fail mainly because the organic dye layer breaks down or, on discs with a silver reflective layer, the metal corrodes, and the public tell-tale is a rising correctable-error counter before data is lost: block error rate (BLER, counted at the first CD decoder) against a specified maximum of 220 per second on CD, and parity-inner errors summed over eight blocks (PI sum 8) against a maximum of 280 on DVD. In the NIST and Library of Congress study of discs bought in 2005, all 7 CD-R products reached an estimated life of more than 30 years at 25 C and 50% relative humidity (5 of 7 more than 45 years), while the 15 DVD products spread from under 15 years (2 products) to more than 45 years (7 products). Canada's conservation institute ranks a gold-layer phthalocyanine CD-R above 100 years and a dye or non-dye single-layer Blu-ray recordable at 5 to 10 years, and says discs recorded with poor technique can fail in as little as 2 to 10 years. These are accelerated-ageing estimates of temperature and humidity effects only, for disc products of 2005 or earlier, so they rank media and set expectations but are not a failure rate for any disc in use today. https://hesela.com/analyses/postmortem-timeline-log-unknown/https://hesela.com/analyses/postmortem-timeline-log-unknown/#publication-postmortem-timeline-log-unknownMon, 12 Oct 2026 00:00:00 GMTOn GitLab.com’s one primary and one hot-standby secondary, the 10 February 2017 postmortem prints two ends for database rows lost on 31 January 2017: the introduction says 17:20 and 00:00 UTC on that date, and the data-loss section says 17:20–23:30 UTC. It says the exact loss is hard to estimate: at least 5,000 projects, 5,000 comments, and roughly 700 users. The incident log archived at 2017-02-02 03:47:19 UTC prints ±6 hours, 5,037 projects, ±4,979 comments, and 707 users that were hard to confirm from Kibana. Those figures are vendor claims, not a census of objects that stayed gone. The same postmortem says projects restored from remaining Git data came back only in basic form, without issues or merge requests. https://hesela.com/analyses/replication-acknowledgements-durability-lag/https://hesela.com/analyses/replication-acknowledgements-durability-lag/#publication-replication-acknowledgements-durability-lagMon, 12 Oct 2026 00:00:00 GMTCompare replication acknowledgements, WAL retention and logical failover readiness; distinguish durable writes, replay and missing or unusable slots.